From e06625d202ba53836a16865e0f779a44a85ec167 Mon Sep 17 00:00:00 2001 From: lsdsjy <1356263+lsdsjy@users.noreply.github.com> Date: Thu, 20 Aug 2026 19:43:15 +0800 Subject: [PATCH 1/2] fix(web): switch running drafts to Send --- ...08-20-running-draft-primary-send.i18n.yaml | 6 +++ .../2026-08-20-running-draft-primary-send.md | 31 +++++++++++ ...026-08-20-running-draft-primary-send.zh.md | 31 +++++++++++ ...8-06-web-queue-steer-all-gesture.i18n.yaml | 4 +- .../2026-08-06-web-queue-steer-all-gesture.md | 2 +- ...26-08-06-web-queue-steer-all-gesture.zh.md | 2 +- apps/web/tests/live-interactions.e2e.ts | 35 ++++++++++--- .../running-draft.expected.md | 28 ++++++++++ .../client/ui-conversation/README.i18n.yaml | 4 +- packages/client/ui-conversation/README.md | 2 + packages/client/ui-conversation/README.zh.md | 2 + .../src/client/skeleton/InputBar.tsx | 8 +-- .../tests/input-bar.client.spec.tsx | 52 +++++++++++++++++-- 13 files changed, 184 insertions(+), 23 deletions(-) create mode 100644 .agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.i18n.yaml create mode 100644 .agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.md create mode 100644 .agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.zh.md create mode 100644 apps/web/tests/snapshots/live-interactions/running-draft.expected.md diff --git a/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.i18n.yaml b/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.i18n.yaml new file mode 100644 index 0000000000..5a1d90d7f1 --- /dev/null +++ b/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.i18n.yaml @@ -0,0 +1,6 @@ +# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each +# side as of the last confirmed-consistent state. Both languages carry equal authority; +# after editing either side, bring the other along and re-record with: +# pnpm run verify-translation-pairing --write .agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.md +2026-08-20-running-draft-primary-send.md: 79c8c3a74fc49a325b229324b0d96b2387308d3d +2026-08-20-running-draft-primary-send.zh.md: 180352810f57ecc04882322318f9708cf247a09c diff --git a/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.md b/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.md new file mode 100644 index 0000000000..79c8c3a74f --- /dev/null +++ b/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.md @@ -0,0 +1,31 @@ +# Agent Note: Running drafts take the primary Send action + +Status: implemented + +English | [中文](2026-08-20-running-draft-primary-send.zh.md) + +## Problem + +The ordinary Web composer remains editable while a Turn is running, and keyboard submission can queue or steer its draft. Its single primary pointer control nevertheless stayed on Stop for the entire Turn. A pointer user who entered a follow-up and activated that control stopped the current Turn instead of submitting the visible draft, so the control contradicted the composer's editable state and the user's current content. + +## Decision + +`InputBar` chooses the ordinary session's primary action from the running state, draft content, and owner block. An empty running composer shows Stop and routes it through the existing session cancellation callback. Non-whitespace text or at least one attachment changes that same control to Send; the click uses the existing Queue submission path. An owner-blocked running composer keeps Stop even when a retained draft exists, because the block disables both editing and submission. Clearing the draft or completing a successful submission restores Stop while the Turn remains active. Idle sessions continue to show Send, disabled while the draft is empty or submission is unavailable. + +The pointer action does not inherit the `ui-conversation.busyEnter` preference. That preference continues to choose Queue or Steer only for the two keyboard gestures. Continuable subagents retain independent Send and Stop controls, and one-shot subagents retain their read-only behavior. + +## Verification + +The `InputBar` component test covers empty, text, cleared, submitted, attachment-only, and owner-blocked running drafts, including Queue submission while the keyboard preference selects Steer. The keyless assembled Web scenario parks a real composed Turn in the replay adapter, captures the running draft with Send, clicks it through the Host Queue path, observes Stop return after the draft clears, removes the queued row, and then cancels the Turn. + +## Alternatives considered + +**Keep Stop for the whole running Turn.** This preserves immediate cancellation but leaves the visible editable draft without a pointer submission action and makes the primary control act against the content beside it. + +**Render Send and Stop simultaneously for every running session.** Continuable subagents need two independent operations because their cancellation route differs from continuation delivery. Ordinary sessions have one established primary seat; adding a permanent second control would spend more space and create a different hierarchy when the draft itself already identifies the immediate action. + +**Apply the busy-Enter preference to pointer Send.** A button labeled Send would silently change between Queue and Steer according to a keyboard preference. Keeping pointer submission on Queue preserves the existing explicit distinction and avoids an invisible mode on the button. + +## Consequences + +Pointer users can submit a follow-up without waiting for the active Turn or using a keyboard shortcut. An actionable draft occupies the single primary seat, so Stop returns after the draft is cleared or accepted rather than remaining simultaneously visible; an owner block returns that seat to Stop because the retained draft cannot be edited or submitted. Keyboard delivery selection, cancellation transport, and subagent controls are unchanged. Issue #2850 records the user-facing defect and acceptance boundary. diff --git a/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.zh.md b/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.zh.md new file mode 100644 index 0000000000..180352810f --- /dev/null +++ b/.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.zh.md @@ -0,0 +1,31 @@ +# Agent Note: 运行中草稿取得主 Send 操作 + +Status: implemented + +[English](2026-08-20-running-draft-primary-send.md) | 中文 + +## 问题 + +普通 Web composer 在 Turn 运行期间仍可编辑，键盘提交也能把草稿送入 Queue 或 Steer。然而，其唯一的主指针控件会在整个 Turn 中一直保持 Stop。指针用户输入后续消息并激活该控件时，会停止当前 Turn，而不是提交眼前的草稿；该控件因而与 composer 的可编辑状态和用户当前内容相冲突。 + +## 决策 + +`InputBar` 根据运行状态、草稿内容和 owner block 选择普通会话的主操作。运行中的 composer 为空时显示 Stop，并通过既有会话取消回调执行。存在非空白文字或至少一个附件时，同一控件切换为 Send，点击后使用既有 Queue 提交路径。owner block 会禁用编辑与提交，因此即使保留了草稿，运行中的 composer 也保持 Stop。清空草稿或成功提交后，只要 Turn 仍在运行，就会恢复 Stop。空闲会话仍显示 Send；草稿为空或无法提交时，该按钮保持禁用。 + +指针操作不继承 `ui-conversation.busyEnter` 偏好。该偏好仍然只为两个键盘手势选择 Queue 或 Steer。可继续 subagent 保留相互独立的 Send 与 Stop 控件，one-shot subagent 保持只读行为。 + +## 验证 + +`InputBar` 组件测试覆盖运行中草稿的空白、文字、清空、提交成功、仅附件和 owner-blocked 状态，并证明键盘偏好选择 Steer 时，按钮提交仍使用 Queue。无密钥的组装 Web 场景通过 replay 适配器停住真实组合出的 Turn，捕获显示 Send 的运行中草稿，经 Host Queue 路径点击提交，在草稿清空后观察 Stop 恢复，移除 Queue 行，再取消该 Turn。 + +## 备选方案 + +**在整个运行中 Turn 保持 Stop。** 这样可以始终立即取消，但可见的可编辑草稿没有指针提交操作，主控件也会执行与相邻内容相反的动作。 + +**为每个运行中会话同时渲染 Send 与 Stop。** 可继续 subagent 需要两个独立操作，因为其取消路由与继续投递不同。普通会话已有单一主操作位置；永久增加第二个控件会占用更多空间，并在草稿本身已经指明当前操作时引入另一套层级。 + +**让指针 Send 采用 busy-Enter 偏好。** 标记为 Send 的按钮会随键盘偏好在 Queue 与 Steer 之间静默变化。保持指针提交始终使用 Queue，可以保留既有显式区分，避免按钮携带不可见模式。 + +## 影响 + +指针用户无需等待当前 Turn 结束或使用键盘快捷键，即可提交后续消息。可操作草稿会占用唯一的主操作位置，因此 Stop 会在草稿清空或被接纳后恢复，而不是同时显示；owner block 会让该位置恢复 Stop，因为保留的草稿无法编辑或提交。键盘投递选择、取消传输和 subagent 控件均不变。Issue #2850 记录用户可见缺陷与验收边界。 diff --git a/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.i18n.yaml b/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.i18n.yaml index b47b635768..e7612e322b 100644 --- a/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.i18n.yaml +++ b/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.md -2026-08-06-web-queue-steer-all-gesture.md: e546f68647dfc9b91ce4699cef4a64694ebc4f76 -2026-08-06-web-queue-steer-all-gesture.zh.md: 1f65933ac430ca22ac8d6c78471a3bd08b421ad6 +2026-08-06-web-queue-steer-all-gesture.md: c51e837d37b61a73f9602445daea9fbdfc77a299 +2026-08-06-web-queue-steer-all-gesture.zh.md: b0009a7a51ebe89f61d1a4b995b1138a7247d266 diff --git a/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.md b/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.md index e546f68647..c51e837d37 100644 --- a/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.md +++ b/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.md @@ -30,4 +30,4 @@ The per-row 插话发送 action and its strict-steer boundary are owned by [Stee - **Steering via `session.prompt(mode: 'steer')` per row.** Rejected: that mints new messages instead of transferring the pending occurrences and would split the dock's immutable-message contract; `updateQueue({ kind: 'steer' })` already atomically transfers the exact occurrence. - **Firing all row steers concurrently.** Rejected: arrival order at the host is not guaranteed, and steering order is model-visible; sequential awaits preserve FIFO. - **A new host RPC for steer-all.** Rejected: the existing per-item operation is idempotent enough — each row is one strict steer, and mid-flush closure converges silently — so a protocol change buys nothing. -- **A send-button tooltip.** Rejected: the primary button is Stop while an ordinary session is running, which is the only window where the whole-queue gesture is available. The empty-draft placeholder occupies that exact window and can describe the keyboard action directly. +- **A send-button tooltip.** Rejected: the primary button is Stop in the empty-draft running window, which is also the only window where the whole-queue gesture is available. The placeholder occupies that exact window and can describe the keyboard action directly. diff --git a/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.zh.md b/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.zh.md index 1f65933ac4..b0009a7a51 100644 --- a/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.zh.md +++ b/.agents/notes/implemented/feature/2026-08-06-web-queue-steer-all-gesture.zh.md @@ -30,4 +30,4 @@ Status: implemented - **逐条用 `session.prompt(mode: 'steer')` 插话。** 已拒绝：那会铸造新消息而不是转移 pending 行，破坏 dock 的不可变消息契约；`updateQueue({ kind: 'steer' })` 已经原子地转移了确切的那条。 - **并发触发所有行。** 已拒绝：host 到达顺序无法保证，而插话顺序对模型可见；顺序 await 保证 FIFO。 - **为 steer-all 新增 host RPC。** 已拒绝：现有逐条操作已足够幂等——每行一次严格 steer，中途关闭静默收敛——协议改动没有收益。 -- **发送按钮 tooltip。** 已拒绝：普通会话运行时，主按钮是 Stop，这也是整队列手势唯一可用的窗口。空草稿时的 placeholder 恰好在该窗口显示，可以直接说明这项键盘操作。 +- **发送按钮 tooltip。** 已拒绝：主按钮在空草稿的运行窗口内是 Stop，而这也正是整队列手势唯一可用的窗口。placeholder 恰好在该窗口显示，可以直接说明这项键盘操作。 diff --git a/apps/web/tests/live-interactions.e2e.ts b/apps/web/tests/live-interactions.e2e.ts index 85a7a57e62..0586a900c2 100644 --- a/apps/web/tests/live-interactions.e2e.ts +++ b/apps/web/tests/live-interactions.e2e.ts @@ -1,6 +1,7 @@ -// Web e2e scenarios: live-turn interactions — cancellation, error surfacing, -// transient-retry recovery, and retry exhaustion, all through the real -// composition and wire. The model adapter is dsh-llm-replay with override +// Web e2e scenarios: live-turn interactions — running-draft submission, +// cancellation, error surfacing, transient-retry recovery, and retry +// exhaustion, all through the real composition and wire. The model adapter is +// dsh-llm-replay with override // sidecars: `hang` (+ a readyFile marker) makes mid-stream cancel // deterministic by construction, `throw` entries express provider failures by // stable code, and `{ patches }` augmentation injects transient throws before @@ -29,11 +30,12 @@ import { connectFreshWorkspace, newEnglishPage, saveFailureShot } from './suppor const SNAPSHOT_DIR = fileURLToPath(new URL('./snapshots/live-interactions', import.meta.url)) const FIXTURE = join(SNAPSHOT_DIR, 'session.jsonl') -// One golden pins the stable mid-turn loading state; the other four capture -// what the user is left looking at after cancel, after a non-retryable failure, -// after retry recovery, and after retry exhaustion. +// One golden pins the empty mid-turn loading state, one pins the sendable draft +// state, and the other four capture what remains after cancel, after a +// non-retryable failure, after retry recovery, and after retry exhaustion. const CANCEL_EXPECTED = join(SNAPSHOT_DIR, 'cancel.expected.md') const LOADING_EXPECTED = join(SNAPSHOT_DIR, 'loading.expected.md') +const RUNNING_DRAFT_EXPECTED = join(SNAPSHOT_DIR, 'running-draft.expected.md') const ERROR_EXPECTED = join(SNAPSHOT_DIR, 'error-auth.expected.md') const RETRY_EXPECTED = join(SNAPSHOT_DIR, 'retry.expected.md') const RETRY_EXHAUSTED_EXPECTED = join(SNAPSHOT_DIR, 'retry-exhausted.expected.md') @@ -44,6 +46,7 @@ const AUTH_PROVIDER_MESSAGE = 'Authentication Fails, Your api key: sk-preview-se // patch. Kept deliberately tool-free so the derived script is exactly one // model call. const PROMPT = 'Reply with a one-sentence description of event sourcing, then stop.' +const RUNNING_DRAFT = 'Queue this follow-up while the current turn is running.' /** turn/end reasons observed, in order. */ function turnEndReasons(events: SessionEvent[]): string[] { @@ -147,6 +150,22 @@ describe('web e2e: live-turn interactions (cancel / error / retry)', () => { ).toBe(true) const loadingSnapshot = await captureStableAria(page, '[class*="centerCol"]', scaffold!.workspaceCwd) await compareOrRefreshGolden(LOADING_EXPECTED, loadingSnapshot, MODE) + + const input = page.locator('textarea').first() + await input.fill(RUNNING_DRAFT) + const send = page.getByRole('button', { name: 'Send message', exact: true }) + await send.waitFor({ timeout: 10_000 }) + expect(await page.getByRole('button', { name: 'Stop generating', exact: true }).count()).toBe(0) + const runningDraftSnapshot = await captureStableAria(page, '[class*="centerCol"]', scaffold!.workspaceCwd) + await compareOrRefreshGolden(RUNNING_DRAFT_EXPECTED, runningDraftSnapshot, MODE) + await send.click() + await expect.poll(() => input.inputValue(), { timeout: 10_000 }).toBe('') + const queuedRow = page.locator('[data-queue-dock]').getByRole('listitem').filter({ hasText: RUNNING_DRAFT }) + await queuedRow.waitFor({ timeout: 10_000 }) + await page.getByRole('button', { name: 'Stop generating', exact: true }).waitFor({ timeout: 10_000 }) + await queuedRow.getByRole('button', { name: 'Remove queued message' }).click() + await expect.poll(() => queuedRow.count(), { timeout: 10_000 }).toBe(0) + await page.getByRole('button', { name: 'Stop generating' }).click() await settled expect(turnEndReasons(sessionEvents).at(-1)).toBe('aborted') @@ -281,8 +300,8 @@ describe('web e2e: live-turn interactions (cancel / error / retry)', () => { it.skipIf(MODE === 'record')('keeps the fixture inventory closed', async () => { await assertFixtureInventory(SNAPSHOT_DIR, [ - 'session.jsonl', 'cancel.expected.md', 'loading.expected.md', 'error-auth.expected.md', 'retry.expected.md', - 'retry-exhausted.expected.md', + 'session.jsonl', 'cancel.expected.md', 'loading.expected.md', 'running-draft.expected.md', + 'error-auth.expected.md', 'retry.expected.md', 'retry-exhausted.expected.md', ]) }) }) diff --git a/apps/web/tests/snapshots/live-interactions/running-draft.expected.md b/apps/web/tests/snapshots/live-interactions/running-draft.expected.md new file mode 100644 index 0000000000..d4e15c0498 --- /dev/null +++ b/apps/web/tests/snapshots/live-interactions/running-draft.expected.md @@ -0,0 +1,28 @@ +- banner: + - navigation "Session hierarchy": + - button "Reply with a one-sentence description" [disabled] + - img + - text: Standard mode + - button "Session log": + - text: Session log + - img + - tablist: + - tab "Chat" [selected] + - tab "Trajectory" +- text: Reply with a one-sentence description of event sourcing, then stop. {{clock}} +- button "Copy": + - img +- button "Context injection @deepseek-ai/dsh-system-prompt": + - img + - img + - text: Context injection @deepseek-ai/dsh-system-prompt +- paragraph: partial +- status: Deep diving... +- textbox "Message the agent": Queue this follow-up while the current turn is running. +- button "Commands": + - img +- 'button "Access mode, current: Workspace Write"': Workspace Write +- button "Select model, current DeepSeek-V4-Flash": + - text: DeepSeek-V4-Flash + - img +- button "Send message" diff --git a/packages/client/ui-conversation/README.i18n.yaml b/packages/client/ui-conversation/README.i18n.yaml index a6102ddac0..85e96fbdca 100644 --- a/packages/client/ui-conversation/README.i18n.yaml +++ b/packages/client/ui-conversation/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/ui-conversation/README.md -README.md: 4c9665b680fe1922770403d88a04dffc755481ad -README.zh.md: 8bf3db2cb1401427f29016f8dbddcd9d27ec9635 +README.md: 08d1408bea404a80da491078aa013989835fccaf +README.zh.md: 874aea8865e6f74a64da2c249087bdad032d98d5 diff --git a/packages/client/ui-conversation/README.md b/packages/client/ui-conversation/README.md index 4c9665b680..08d1408bea 100644 --- a/packages/client/ui-conversation/README.md +++ b/packages/client/ui-conversation/README.md @@ -20,6 +20,8 @@ View selection is deterministic: a registered persisted selection wins, otherwis The resident composer survives no-Session and Session transitions. The no-Session state keeps the same textarea mounted but inert while the Workspace picker connects a blank Session. Draft text is mirrored into the per-Session Conversation store. Queue operations address exact queue occurrences through the scoped `ctx.conversation` service. Busy Enter behavior is stored in the Host-backed `ui-conversation` settings namespace. +An ordinary running composer keeps Stop as its primary pointer action while its draft is empty or an owner block makes input unavailable. Actionable text or attachments switch the same seat to Queue Send; clearing or successfully submitting the draft restores Stop. Keyboard Queue/Steer selection remains governed by the busy-Enter setting, while continuable subagents keep independent Send and Stop actions ([decision](../../../.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.md)). + ## Temporary composer entries `conversation.composer` is a generic chain. Its complete owner currency is: diff --git a/packages/client/ui-conversation/README.zh.md b/packages/client/ui-conversation/README.zh.md index 8bf3db2cb1..874aea8865 100644 --- a/packages/client/ui-conversation/README.zh.md +++ b/packages/client/ui-conversation/README.zh.md @@ -20,6 +20,8 @@ View 选择规则固定：有效且已注册的持久化选择优先，其次是 常驻 composer 在无 Session 与有 Session 之间保持挂载。无 Session 时，同一个 textarea 保持 inert，Workspace picker 连接 blank Session；草稿文本镜像到逐 Session Conversation store。Queue 操作通过 scoped `ctx.conversation` service 寻址准确的 queue occurrence。繁忙时 Enter 行为保存在 Host-backed `ui-conversation` settings namespace。 +普通 composer 运行期间，草稿为空或 owner block 使输入不可用时，主指针操作保持为 Stop。可提交文字或附件会把同一位置切换为 Queue Send；清空或成功提交草稿后恢复 Stop。键盘 Queue/Steer 选择仍由繁忙态 Enter 设置决定，可继续 subagent 则保留相互独立的 Send 与 Stop 操作（[决策](../../../.agents/notes/implemented/bug-fix/2026-08-20-running-draft-primary-send.zh.md)）。 + ## 临时 composer entry `conversation.composer` 是通用 chain，其完整 owner currency 为： diff --git a/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx b/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx index 8bedb2d260..0a137c8010 100644 --- a/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx +++ b/packages/client/ui-conversation/src/client/skeleton/InputBar.tsx @@ -554,10 +554,10 @@ export function InputBar({ if (el !== null) toggleCommandMenu?.(selectionOf(el)) } - // Ordinary sessions retain their primary Send/Stop toggle. A continuable - // child keeps Send as the primary action and exposes Stop independently so - // pointer users can queue follow-ups while its current turn is running. - const primaryStops = running && subagent === null + // An ordinary running session keeps Stop while the composer is empty or + // owner-blocked; an actionable draft gets the existing Queue action. A + // continuable child keeps Send primary and exposes Stop independently. + const primaryStops = running && subagent === null && (empty || blocked !== undefined) const interruptible = running && continuable const primaryLabel = primaryStops ? t('input.stop') : t('input.send') const onPrimary = (): void => { diff --git a/packages/client/ui-conversation/tests/input-bar.client.spec.tsx b/packages/client/ui-conversation/tests/input-bar.client.spec.tsx index f36b1aa1c4..17064a97a5 100644 --- a/packages/client/ui-conversation/tests/input-bar.client.spec.tsx +++ b/packages/client/ui-conversation/tests/input-bar.client.spec.tsx @@ -64,6 +64,7 @@ interface BenchOptions { subagent?: Exclude disabled?: boolean inert?: boolean + blocked?: { readonly reason: string } workspacePickerOpen?: boolean onRequestWorkspace?: () => void promptError?: SessionSnapshot['promptError'] @@ -186,6 +187,7 @@ function bench(over?: BenchOptions) { renderSlot, variant: over?.variant ?? 'composer', ...(over?.inert === true ? { disabled: true } : {}), + ...(over?.blocked !== undefined ? { blocked: over.blocked } : {}), ...(over?.workspacePickerOpen !== undefined ? { workspacePickerOpen: over.workspacePickerOpen } : {}), ...(over?.onRequestWorkspace !== undefined ? { onRequestWorkspace: over.onRequestWorkspace } : {}), ...(over?.placeholder !== undefined ? { placeholder: over.placeholder } : {}), @@ -196,7 +198,9 @@ function bench(over?: BenchOptions) { } const view = render() const textarea = view.container.querySelector('textarea')! + const sendableDraft = (over?.draft?.trim() ?? '') !== '' || (over?.attachments?.length ?? 0) > 0 const primaryStops = over?.running === true && over.subagent === undefined + && (!sendableDraft || over.blocked !== undefined) const button = view.container.querySelector( `button[aria-label="${primaryStops ? '停止生成' : '发送消息'}"]`, )! @@ -606,15 +610,53 @@ describe('Enter semantics', () => { }) describe('running and lock semantics', () => { - it('running keeps the input free (typing + Enter queue) while the primary turns stop', () => { - const { textarea, button, stop, sink } = bench({ running: true, draft: '排队消息' }) + it('running switches the primary between Stop and Queue Send with the draft', async () => { + const { textarea, button, stop, sink } = bench({ running: true, busyEnter: 'steer' }) expect(textarea.disabled).toBe(false) - fireEvent.change(textarea, { target: { value: '排队消息2' } }) - fireEvent.keyDown(textarea, { key: 'Enter' }) - expect(sink).toHaveBeenCalledWith('排队消息2', [], 'queue', expect.any(AbortSignal)) expect(button.getAttribute('aria-label')).toBe('停止生成') fireEvent.click(button) expect(stop).toHaveBeenCalledTimes(1) + + fireEvent.change(textarea, { target: { value: '排队消息' } }) + expect(button.getAttribute('aria-label')).toBe('发送消息') + fireEvent.change(textarea, { target: { value: ' ' } }) + expect(button.getAttribute('aria-label')).toBe('停止生成') + fireEvent.change(textarea, { target: { value: '排队消息2' } }) + expect(button.getAttribute('aria-label')).toBe('发送消息') + fireEvent.click(button) + expect(sink).toHaveBeenCalledWith('排队消息2', [], 'queue', expect.any(AbortSignal)) + await vi.waitFor(() => { expect(button.getAttribute('aria-label')).toBe('停止生成') }) + expect(stop).toHaveBeenCalledTimes(1) + }) + + it('running treats an attachment-only draft as Send', async () => { + const attachment = { + kind: 'image' as const, + id: 'draft-1' as DraftAttachmentId, + file: new File([Uint8Array.of(1)], 'pixel.png', { type: 'image/png' }), + previewUrl: 'blob:pixel', + } + const { button, sink } = bench({ running: true, attachments: [attachment] }) + expect(button.getAttribute('aria-label')).toBe('发送消息') + fireEvent.click(button) + expect(sink).toHaveBeenCalledWith('', ['draft-1'], 'queue', expect.any(AbortSignal)) + await vi.waitFor(() => { expect(button.getAttribute('aria-label')).toBe('停止生成') }) + }) + + it('running blocked composer keeps Stop with a retained draft', () => { + const { button, sink, stop, textarea } = bench({ + running: true, + draft: '保留的草稿', + blocked: { reason: '请选择可用模型' }, + placeholder: '请选择可用模型', + }) + expect(textarea.disabled).toBe(true) + expect(textarea.placeholder).toBe('请选择可用模型') + expect(button.getAttribute('aria-label')).toBe('停止生成') + expect(button.disabled).toBe(false) + fireEvent.click(button) + expect(stop).toHaveBeenCalledTimes(1) + expect(sink).not.toHaveBeenCalled() }) it('running plain Enter follows the busy-state Steer preference', () => { From f76a225a7db1560e1ed8b77d30fe4f2e7b774d65 Mon Sep 17 00:00:00 2001 From: Dudu-0223 Date: Mon, 24 Aug 2026 18:23:42 +0800 Subject: [PATCH 2/2] Merge pull request #2663 from deepseek-harness/feat/subagent-provider MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 让 subagent 按需发现并选择子 Agent 模型 --- ...8-10-fork-children-stay-one-shot.i18n.yaml | 4 +- .../2026-08-10-fork-children-stay-one-shot.md | 20 +- ...26-08-10-fork-children-stay-one-shot.zh.md | 20 +- ...6-06-21-subagent-capability-seam.i18n.yaml | 4 +- .../2026-06-21-subagent-capability-seam.md | 6 +- .../2026-06-21-subagent-capability-seam.zh.md | 6 +- ...8-model-selected-subagent-routes.i18n.yaml | 6 + ...26-08-18-model-selected-subagent-routes.md | 61 ++ ...08-18-model-selected-subagent-routes.zh.md | 61 ++ apps/cli/tests/web-agent-presets.e2e.ts | 29 + .../models-settings/configured.expected.md | 4 + .../models-settings/declared-edit.expected.md | 4 + .../models-settings/declared.expected.md | 4 + .../models-settings/empty.expected.md | 4 + .../models.expected.md | 4 + .../dismissed.expected.md | 4 + docs/capability-seams.i18n.yaml | 4 +- docs/capability-seams.md | 6 +- docs/capability-seams.zh.md | 6 +- docs/config-catalog.i18n.yaml | 4 +- docs/config-catalog.md | 12 +- docs/config-catalog.zh.md | 10 +- docs/event-producer-consumer.i18n.yaml | 4 +- docs/event-producer-consumer.md | 34 +- docs/event-producer-consumer.zh.md | 16 +- docs/module-graph.i18n.yaml | 4 +- docs/module-graph.md | 336 +++++------ docs/module-graph.zh.md | 338 +++++------ docs/persistence-catalog.i18n.yaml | 4 +- docs/persistence-catalog.md | 18 +- docs/persistence-catalog.zh.md | 18 +- docs/subsystems/core.i18n.yaml | 4 +- docs/subsystems/core.md | 8 +- docs/subsystems/core.zh.md | 8 +- docs/subsystems/subagent.i18n.yaml | 4 +- docs/subsystems/subagent.md | 27 +- docs/subsystems/subagent.zh.md | 27 +- docs/tool-catalog.i18n.yaml | 4 +- docs/tool-catalog.md | 40 +- docs/tool-catalog.zh.md | 40 +- examples/acp-agent/cordis.yml | 8 + .../acp-agent/depth-two.cordis.snapshot.yml | 1 + examples/acp-agent/depth-two.cordis.yml | 1 + ...gent-configured-effort.cordis.snapshot.yml | 65 +++ .../subagent-configured-effort.cordis.yml | 15 + examples/acp-agent/tests/acp.snapshot.ts | 25 +- .../advanced-toolchain/session.1.jsonl | 4 +- .../advanced-toolchain/session.2.jsonl | 4 +- .../system-prompt.expected.md | 16 +- .../tool-schemas.expected.json | 31 +- .../both-mode-turn/system-prompt.expected.md | 16 +- .../both-mode-turn/tool-schemas.expected.json | 31 +- .../system-prompt.expected.md | 16 +- .../code-mode-turn/system-prompt.expected.md | 16 +- .../tool-schemas.expected.json | 31 +- .../lsp-definition/tool-schemas.expected.json | 31 +- .../tool-schemas.expected.json | 31 +- .../tool-schemas.expected.json | 31 +- .../tool-schemas.expected.json | 31 +- .../pty-tools/tool-schemas.expected.json | 31 +- .../tool-schemas.expected.json | 31 +- .../session.1.jsonl | 2 +- .../tool-schemas.expected.json | 31 +- .../input.json | 7 + .../replay.override.json | 32 + .../session.jsonl | 41 ++ .../stdout.expected.jsonl | 8 + .../session.1.jsonl | 2 +- .../tool-schemas.1.expected.json | 31 +- .../subagent-continuable/session.1.jsonl | 2 +- .../tool-schemas.1.expected.json | 31 +- .../session.1.jsonl | 4 +- .../session.2.jsonl | 4 +- .../subagent-fork-in-process/session.1.jsonl | 2 +- .../subagent-list-agents/session.1.jsonl | 2 +- .../tool-schemas.1.expected.json | 31 +- .../session.1.jsonl | 2 +- .../snapshots/subagent-mixed/session.1.jsonl | 4 +- .../snapshots/subagent-mixed/session.2.jsonl | 4 +- .../snapshots/subagent-multi/session.1.jsonl | 4 +- .../snapshots/subagent-multi/session.2.jsonl | 4 +- .../subagent-parallel/session.1.jsonl | 10 +- .../subagent-parallel/session.2.jsonl | 10 +- .../snapshots/subagent-report/session.1.jsonl | 2 +- .../tool-schemas.1.expected.json | 31 +- .../subagent-spawn-in-process/session.1.jsonl | 2 +- .../text-turn/tool-schemas.expected.json | 31 +- .../web-fetch/tool-schemas.expected.json | 31 +- .../snapshots/workflow-run/session.1.jsonl | 2 +- examples/headless-agent/cordis.yml | 12 +- .../advanced-toolchain/session.1.jsonl | 12 +- .../advanced-toolchain/session.2.jsonl | 12 +- .../advanced-toolchain/session.jsonl | 28 +- .../compaction-recovery/session.jsonl | 14 +- .../tests/snapshots/pty-tools/session.jsonl | 34 +- .../subagent-settlement/child.expected.jsonl | 2 +- .../parent-override/child.expected.jsonl | 4 +- .../parent-override/parent.expected.jsonl | 9 +- .../tests/subagent-inheritance.snapshot.ts | 21 +- examples/python-sdk-agent/cordis.yml | 1 + .../tests/keyless-smoke.e2e.ts | 2 + .../notifications.expected.jsonl | 2 +- .../subagent-spawn-in-process/session.1.jsonl | 2 +- .../tests/session-cold.host.spec.ts | 7 +- packages/bundle/base/cordis.patch.yml | 11 +- packages/bundle/web-app/cordis.patch.yml | 5 + packages/bundle/web-app/package.json | 1 + .../ui-settings-models/README.i18n.yaml | 4 +- packages/client/ui-settings-models/README.md | 2 + .../client/ui-settings-models/README.zh.md | 2 + .../src/client/ModelsSection.module.css | 84 ++- .../src/client/ModelsSection.tsx | 13 + .../src/client/SubagentModelSelectionCard.tsx | 87 +++ .../ui-settings-models/src/client/locales.ts | 8 + .../ui-settings-models/src/client/store.ts | 9 + .../tests/components.client.spec.tsx | 81 ++- packages/core/agent-loop/README.i18n.yaml | 4 +- packages/core/agent-loop/README.md | 3 +- packages/core/agent-loop/README.zh.md | 3 +- packages/core/agent-loop/src/agent.ts | 3 +- packages/core/agent-loop/src/index.ts | 3 +- packages/core/agent-loop/tests/loop.spec.ts | 41 +- packages/core/agent/README.i18n.yaml | 4 +- packages/core/agent/README.md | 2 +- packages/core/agent/README.zh.md | 2 +- packages/core/agent/src/runtime-types.ts | 4 +- .../core/session/src/known-event-types.ts | 1 + .../core/tools/tests/gen-tool-catalog.spec.ts | 4 +- .../preview-architecture-review/session.jsonl | 2 +- .../preview-follow-up-builder/session.jsonl | 2 +- .../extensions/tool-cordis/src/api-catalog.ts | 21 +- packages/llm/llm-deepseek/README.i18n.yaml | 4 +- packages/llm/llm-deepseek/README.md | 4 +- packages/llm/llm-deepseek/README.zh.md | 4 +- packages/llm/llm-deepseek/src/adapter.ts | 30 +- packages/llm/llm-deepseek/src/index.ts | 14 +- .../llm/llm-deepseek/tests/adapter.spec.ts | 66 ++- .../preset/agent-presets/README.i18n.yaml | 4 +- packages/preset/agent-presets/README.md | 2 +- packages/preset/agent-presets/README.zh.md | 2 +- packages/preset/agent-presets/package.json | 1 + .../presets/code/agent.cordis.yml | 5 + .../presets/cordis/agent.cordis.yml | 5 + .../presets/standard/agent.cordis.yml | 5 + packages/preset/agent-presets/src/index.ts | 14 +- .../preset/agent-presets/tests/mount.spec.ts | 21 + packages/preset/agent-presets/tsconfig.json | 3 + .../server/tests/built-scope-carrier.e2e.ts | 2 +- packages/sdk/server/tests/server.spec.ts | 10 +- .../subagent/subagent-acp/README.i18n.yaml | 4 +- packages/subagent/subagent-acp/README.md | 6 +- packages/subagent/subagent-acp/README.zh.md | 6 +- packages/subagent/subagent-acp/src/index.ts | 12 +- .../subagent-acp/tests/subagent-acp.spec.ts | 8 +- .../subagent-claude-code/README.i18n.yaml | 4 +- .../subagent/subagent-claude-code/README.md | 4 +- .../subagent-claude-code/README.zh.md | 4 +- .../tests/loader-composition.e2e.ts | 4 + .../subagent/subagent-codex/README.i18n.yaml | 4 +- packages/subagent/subagent-codex/README.md | 4 +- packages/subagent/subagent-codex/README.zh.md | 4 +- .../tests/loader-composition.e2e.ts | 3 + .../subagent-dsh-sdk/README.i18n.yaml | 4 +- packages/subagent/subagent-dsh-sdk/README.md | 6 +- .../subagent/subagent-dsh-sdk/README.zh.md | 6 +- .../subagent/subagent-dsh-sdk/src/index.ts | 2 +- .../tests/subagent-dsh-sdk.spec.ts | 1 + .../subagent-fork-in-process/README.i18n.yaml | 4 +- .../subagent-fork-in-process/README.md | 7 +- .../subagent-fork-in-process/README.zh.md | 7 +- .../subagent-fork-in-process/src/index.ts | 23 +- .../tests/subagent-fork-in-process.spec.ts | 10 +- .../README.i18n.yaml | 4 +- .../subagent-in-process-driver/README.md | 2 +- .../subagent-in-process-driver/README.zh.md | 2 +- .../tests/structured.spec.ts | 2 +- .../README.i18n.yaml | 4 +- .../subagent-spawn-in-process/README.md | 8 +- .../subagent-spawn-in-process/README.zh.md | 8 +- .../subagent-spawn-in-process/src/index.ts | 14 +- .../tests/subagent-spawn-in-process.spec.ts | 10 +- packages/subagent/subagent/README.i18n.yaml | 4 +- packages/subagent/subagent/README.md | 7 +- packages/subagent/subagent/README.zh.md | 7 +- packages/subagent/subagent/src/child-agent.ts | 49 +- .../subagent/subagent/src/continuation.ts | 14 +- packages/subagent/subagent/src/descriptor.ts | 11 +- packages/subagent/subagent/src/index.ts | 2 + .../subagent/subagent/src/out-of-process.ts | 3 +- packages/subagent/subagent/src/types.ts | 7 + .../subagent/tests/child-agent.spec.ts | 76 +++ .../subagent/tests/continuation.spec.ts | 66 ++- .../subagent/subagent/tests/invariant.spec.ts | 2 +- .../subagent/tests/out-of-process.spec.ts | 8 +- .../subagent/subagent/tests/service.spec.ts | 16 +- .../subagent/tool-subagent/README.i18n.yaml | 4 +- packages/subagent/tool-subagent/README.md | 33 +- packages/subagent/tool-subagent/README.zh.md | 33 +- packages/subagent/tool-subagent/package.json | 10 + packages/subagent/tool-subagent/src/index.ts | 545 ++++++++++++------ .../subagent/tool-subagent/src/invariant.ts | 26 +- .../subagent/tool-subagent/src/list-models.ts | 94 +++ .../src/model-selection-settings.ts | 70 +++ .../src/model-selection-state.ts | 33 ++ .../tool-subagent/src/model-selection.ts | 112 ++++ .../subagent/tool-subagent/tests/harness.ts | 57 ++ .../tool-subagent/tests/list-models.spec.ts | 191 ++++++ .../tests/model-selection-settings.spec.ts | 248 ++++++++ .../tests/model-selection.spec.ts | 365 ++++++++++++ .../tool-subagent/tests/scripted-provider.ts | 1 + .../tool-subagent/tests/tool-subagent.spec.ts | 169 +++--- packages/subagent/tool-subagent/tsconfig.json | 9 + .../subagent/tool-subagent/tsdown.config.ts | 19 + .../tool-ralph/tests/tool-ralph.spec.ts | 1 + .../tool-workflow/tests/tool-workflow.spec.ts | 2 +- .../tests/built-worker.e2e.ts | 2 +- .../tests/source-worker.compat.spec.ts | 2 +- .../tests/workflow-worker-thread.spec.ts | 20 +- pnpm-lock.yaml | 9 + scripts/check-workspace-constraints.ts | 3 + scripts/gen-cordis-catalog.ts | 1 + scripts/gen-doc-graphs.ts | 8 + scripts/gen-tool-catalog.ts | 15 +- .../advanced/result.json | 14 +- .../advanced/session.1.jsonl | 4 +- .../advanced/session.2.jsonl | 4 +- tsconfig.base.json | 1 + 227 files changed, 4353 insertions(+), 1029 deletions(-) create mode 100644 .agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.i18n.yaml create mode 100644 .agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md create mode 100644 .agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.zh.md create mode 100644 examples/acp-agent/subagent-configured-effort.cordis.snapshot.yml create mode 100644 examples/acp-agent/subagent-configured-effort.cordis.yml create mode 100644 examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/input.json create mode 100644 examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/replay.override.json create mode 100644 examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/session.jsonl create mode 100644 examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/stdout.expected.jsonl create mode 100644 packages/client/ui-settings-models/src/client/SubagentModelSelectionCard.tsx create mode 100644 packages/subagent/subagent/tests/child-agent.spec.ts create mode 100644 packages/subagent/tool-subagent/src/list-models.ts create mode 100644 packages/subagent/tool-subagent/src/model-selection-settings.ts create mode 100644 packages/subagent/tool-subagent/src/model-selection-state.ts create mode 100644 packages/subagent/tool-subagent/src/model-selection.ts create mode 100644 packages/subagent/tool-subagent/tests/harness.ts create mode 100644 packages/subagent/tool-subagent/tests/list-models.spec.ts create mode 100644 packages/subagent/tool-subagent/tests/model-selection-settings.spec.ts create mode 100644 packages/subagent/tool-subagent/tests/model-selection.spec.ts create mode 100644 packages/subagent/tool-subagent/tsdown.config.ts diff --git a/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.i18n.yaml b/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.i18n.yaml index 7c2052c0a3..7342dd6abc 100644 --- a/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.i18n.yaml +++ b/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md -2026-08-10-fork-children-stay-one-shot.md: 44b947a3e0580263f1973aaf24534b7b2f01c0b6 -2026-08-10-fork-children-stay-one-shot.zh.md: acb12c54fa37d4462cac1b1035bc74d5a96ea719 +2026-08-10-fork-children-stay-one-shot.md: b2d9a77d7cc9969e517a4f5d6973aa2aee1134f5 +2026-08-10-fork-children-stay-one-shot.zh.md: b5dc1a7fda4e2b4152baf63e9c49b89bcebdeef0 diff --git a/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md b/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md index 44b947a3e0..b2d9a77d7c 100644 --- a/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md +++ b/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md @@ -1,4 +1,4 @@ -# Agent Note: Forked children stay one-shot +# Agent Note: Cache-preserving forked children stay one-shot Status: implemented @@ -12,7 +12,7 @@ The child-scoped `report` return channel is now the largest such addition, and s ## Decision -Every shipped composition binds the fork delegation tool to `backgroundMode: one-shot`: [the base bundle](../../../../packages/bundle/base/cordis.patch.yml), [the ACP example](../../../../examples/acp-agent/cordis.yml), and [the headless example](../../../../examples/headless-agent/cordis.yml). The base bundle leaves `run_in_background` available, because it mounts a task service; the two examples set `enableRunInBackground: false`, because they mount none and a one-shot background start would otherwise fail at call time on a missing `tasks` service. +The cache-preserving compositions bind the fork delegation tool to `backgroundMode: one-shot`: [the base bundle](../../../../packages/bundle/base/cordis.patch.yml), [the ACP example](../../../../examples/acp-agent/cordis.yml), and [the headless example](../../../../examples/headless-agent/cordis.yml). The base bundle leaves `run_in_background` available, because it mounts a task service; the two examples set `enableRunInBackground: false`, because they mount none and a one-shot background start would otherwise fail at call time on a missing `tasks` service. The standard, code, and Cordis CLI presets instead bind fork to `continuable`; their child-scoped `report` additions invalidate the inherited prefix and accept the recomputation cost described here. One-shot children — foreground and background alike — are created through `SubagentRuntime.start()`, which never enters the continuable activation-setup registry, so neither `report` nor its prompt section is installed. A forked one-shot child's system prompt and tool schemas therefore equal its parent's, apart from the `persona` and `toolFilter` deltas a deployment opts into per delegation tool. @@ -20,9 +20,9 @@ One-shot children — foreground and background alike — are created through `S ### The restriction is composition, not code -`ForkInProcessProvider.prepareContinuable` stays implemented and `ctx.subagents.startContinuable()` still accepts `fork`; only the shipped `cordis.yml` rows changed. `tool-subagent` knows both the provider's `inheritsParentContext` and its own `backgroundMode` at mount, so a load-time rejection of the pair was available and is deliberately not added: the pair is not wrong in general. It is wrong only while a child-scope delta precedes inherited history, and the package that creates that delta — [`dsh-tool-subagent-report`](../../../../packages/subagent/tool-subagent-report/README.md) — is separately installable and, by its own design, invisible to `tool-subagent`. A deployment that omits the report package can run continuable forked children with the prefix intact. Encoding one roster's consequence as a delegation-tool invariant would make the tool assert something it cannot observe. +`ForkInProcessProvider.prepareContinuable` stays implemented and `ctx.subagents.startContinuable()` accepts `fork`; composition chooses whether the fork tool is one-shot or continuable. `tool-subagent` knows both the provider's `inheritsParentContext` and its own `backgroundMode` at mount, so a load-time rejection of the pair is available and deliberately absent: the pair is not wrong in general. It is costly only while a child-scope delta precedes inherited history, and the package that creates that delta — [`dsh-tool-subagent-report`](../../../../packages/subagent/tool-subagent-report/README.md) — is separately installable and, by its own design, invisible to `tool-subagent`. A deployment that omits the report package can run continuable forked children with the prefix intact. Encoding one roster's consequence as a delegation-tool invariant would make the tool assert something it cannot observe. -The reintroduction condition is recorded as a `TODO(fork-continuable-prefix-reuse)` marker on `prepareContinuable` itself, the one method the shipped compositions do not call, and tracked as issue #2124: continuable fork reopens when a child's system prompt and tool schemas can match its parent's byte for byte. +The cache-preserving condition is recorded as a `TODO(fork-continuable-prefix-reuse)` marker on `prepareContinuable` and tracked as issue #2124: continuable fork preserves its inherited prefix when the child's system prompt and tool schemas can match the parent's byte for byte. ## Alternatives considered @@ -30,7 +30,7 @@ The reintroduction condition is recorded as a `TODO(fork-continuable-prefix-reus **Stop mounting the fork provider at all.** This was the broader form of the restriction. Rejected because foreground fork *is* the prefix-reusing case and is untouched by the report channel, so a full ban gives up the capability without buying anything the one-shot binding does not already buy — and would leave no shipped composition exercising session seeding. -**Ship continuable forked children and accept the loss.** Rejected because the loss is total rather than marginal: reuse breaks ahead of the inherited history, so the child pays full prefill on a transcript it duplicated for the sole purpose of not paying it. A deployment that wants a long-lived child with no inherited context already has `spawn`. +**Use continuable forked children in cache-preserving compositions and accept the loss.** Rejected for the base bundle and ACP/headless examples because the loss is total rather than marginal: reuse breaks ahead of the inherited history, so the child pays full prefill on a transcript it duplicated for the sole purpose of not paying it. The CLI presets make the other tradeoff and retain continuable fork. A deployment that wants a long-lived child with no inherited context already has `spawn`. **Make `report` visible to every Agent.** A global registration would restore byte-identical prefixes by giving parent and child the same schema and section. Rejected because roots, one-shot children, remote children, and agentless callers would advertise a tool with no derivable recipient, and execution-time rejection would make schema visibility disagree with authority — the scope-local decision the [report tool Agent Note](../feature/2026-07-30-continuable-subagent-report-tool.md) already settled. @@ -38,12 +38,12 @@ The reintroduction condition is recorded as a `TODO(fork-continuable-prefix-reus ## Consequences -- No shipped composition creates a continuable forked child; `subagent_fork` returns a result to its caller's turn, and `send_message` addresses only spawned children. -- A forked child's request prefix stays byte-identical to its parent's unless the deployment configures `persona` or `toolFilter` on the fork delegation tool, so the token cost of seeding buys provider-side reuse again. -- The fork provider's continuable path has no production caller and no assembled-composition coverage. It keeps its package-level tests, and the seam still accepts it, so a bundle or `--patch` overlay can reintroduce it with no code change and no warning. +- The base bundle and ACP/headless examples create only one-shot forked children; their `subagent_fork` returns a result to the caller's turn, and `send_message` addresses only spawned children there. The three CLI presets create continuable forked children. +- A one-shot forked child's request prefix stays byte-identical to its parent's unless the deployment configures `persona`, `toolFilter`, or a different LLM route on the fork delegation tool, so the token cost of seeding can buy provider-side reuse. Continuable fork adds `report` before the inherited history and forfeits that reuse. +- The fork provider's continuable path has CLI production callers and package-level tests. The same seam accepts one-shot composition, so a bundle or `--patch` overlay can choose either lifecycle without a code change or warning. - `subagent_fork`'s model-visible schema changes: the continuable background wording is replaced by the one-shot task wording in the base bundle, and disappears entirely from the two examples. The affected keyless snapshot tool-schema sidecars are re-recorded in the same change. -- The report obligation's reach narrows to spawned children in shipped deployments. Its default `next-step` scheduling, authority model, and coverage remain independent of fork composition. +- The report obligation reaches spawned children in every continuable composition and forked children in the CLI presets. Its default `next-step` scheduling, authority model, and coverage remain independent of fork composition. ### Accepted risks -The constraint lives in three configuration files and a code comment, not in a gate. A future bundle row or profile patch can set `backgroundMode: continuable` on a fork tool and silently reintroduce the prefix loss; nothing fails loud. That is the accepted cost of not encoding one roster's consequence into `tool-subagent`. +The one-shot constraint lives in three configuration files and a code comment, not in a gate; the CLI preset rows already choose `backgroundMode: continuable` and incur the prefix loss. Any bundle or profile patch can make either choice without a warning. That is the accepted cost of not encoding one roster's consequence into `tool-subagent`. diff --git a/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.zh.md b/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.zh.md index acb12c54fa..b5dc1a7fda 100644 --- a/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.zh.md +++ b/.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.zh.md @@ -1,4 +1,4 @@ -# Agent Note: fork 出的 child 保持 one-shot +# Agent Note: 保留缓存的 fork child 保持 one-shot Status: implemented @@ -12,7 +12,7 @@ fork 与 spawn 的唯一区别是 child 的 Session 会以 parent 已完成轮 ## 决策 -所有随附组合都把 fork 委派工具绑定为 `backgroundMode: one-shot`：[base 组合包](../../../../packages/bundle/base/cordis.patch.yml)、[ACP 示例](../../../../examples/acp-agent/cordis.yml)与[headless 示例](../../../../examples/headless-agent/cordis.yml)。base 组合包保留 `run_in_background`，因为它挂载了 task 服务；两个示例设置 `enableRunInBackground: false`，因为它们都不挂载 task 服务，否则一次 one-shot 后台启动会在调用时因缺少 `tasks` 服务而失败。 +保留缓存的组合会把 fork 委派工具绑定为 `backgroundMode: one-shot`：[base 组合包](../../../../packages/bundle/base/cordis.patch.yml)、[ACP 示例](../../../../examples/acp-agent/cordis.yml)与[headless 示例](../../../../examples/headless-agent/cordis.yml)。base 组合包保留 `run_in_background`，因为它挂载了 task 服务；两个示例设置 `enableRunInBackground: false`，因为它们都不挂载 task 服务，否则一次 one-shot 后台启动会在调用时因缺少 `tasks` 服务而失败。standard、code 与 Cordis CLI preset 则把 fork 绑定为 `continuable`；其子级作用域的 `report` 增量会使继承前缀失效，并接受这里说明的重算成本。 one-shot child——前台与后台皆然——经由 `SubagentRuntime.start()` 创建，该路径从不进入可继续的 activation setup 注册表，因此 `report` 与它的提示词 section 都不会被安装。于是一个 fork 出的 one-shot child 的系统提示词与工具 schema 与其 parent 相同，只差部署逐个委派工具主动选择的 `persona` 与 `toolFilter` 增量。 @@ -20,9 +20,9 @@ one-shot child——前台与后台皆然——经由 `SubagentRuntime.start()` ### 该限制在于组合，不在于代码 -`ForkInProcessProvider.prepareContinuable` 仍然实现完好，`ctx.subagents.startContinuable()` 也仍接受 `fork`；改动的只有随附的 `cordis.yml` 行。`tool-subagent` 在挂载时同时知道提供方的 `inheritsParentContext` 与自身的 `backgroundMode`，因此一个加载期拒绝该组合的检查是可行的，而这里刻意不加：该组合并非普遍错误。它只在某个 child 作用域增量位于继承历史之前时才是错的，而产生该增量的包——[`dsh-tool-subagent-report`](../../../../packages/subagent/tool-subagent-report/README.zh.md)——是独立安装的，并且按其自身设计对 `tool-subagent` 不可见。一个不安装 report 包的部署可以在前缀完好的前提下运行可继续的 fork child。把某一份插件清单的后果写成委派工具的不变量，会让该工具断言它无法观察到的事实。 +`ForkInProcessProvider.prepareContinuable` 仍然实现完好，`ctx.subagents.startContinuable()` 也接受 `fork`；组合会选择 fork 工具采用 one-shot 还是 continuable。`tool-subagent` 在挂载时同时知道提供方的 `inheritsParentContext` 与自身的 `backgroundMode`，因此一个加载期拒绝该组合的检查是可行的，而这里刻意不加：该组合并非普遍错误。只有在某个 child 作用域增量位于继承历史之前时，它才会产生高昂成本，而产生该增量的包——[`dsh-tool-subagent-report`](../../../../packages/subagent/tool-subagent-report/README.zh.md)——是独立安装的，并且按其自身设计对 `tool-subagent` 不可见。一个不安装 report 包的部署可以在前缀完好的前提下运行可继续的 fork child。把某一份插件清单的后果写成委派工具的不变量，会让该工具断言它无法观察到的事实。 -重新开放的条件记录为 `prepareContinuable` 方法上的 `TODO(fork-continuable-prefix-reuse)` 标记——随附组合不调用这个方法——并由 issue #2124 跟踪：当 child 的系统提示词与工具 schema 能与其 parent 逐字节一致时，可继续 fork 即可重新开放。 +保留缓存的条件记录为 `prepareContinuable` 方法上的 `TODO(fork-continuable-prefix-reuse)` 标记，并由 issue #2124 跟踪：当 child 的系统提示词与工具 schema 能与其 parent 逐字节一致时，可继续 fork 就能保留继承前缀。 ## 备选方案 @@ -30,7 +30,7 @@ one-shot child——前台与后台皆然——经由 `SubagentRuntime.start()` **干脆不挂载 fork 提供方。** 这是该限制更彻底的形式。否决的原因是前台 fork *正是*复用前缀的那种情形，且不受 report 通道影响，因此全面禁用会在不换来任何 one-shot 绑定尚未换来的东西的同时放弃该能力——并且随附组合将没有任何一个演练 session 初始内容。 -**照常随附可继续的 fork child 并接受这份损失。** 否决的原因是这份损失是全额而非边际的：复用在继承历史之前就已中断，于是 child 为一份自己复制过来、目的恰恰是不必付费的 transcript 付了全额预填充。想要一个没有继承上下文的长期 child 的部署，本来就有 `spawn`。 +**在保留缓存的组合中随附可继续的 fork child 并接受这份损失。** base 组合包与 ACP/headless 示例不采用，因为这份损失是全额而非边际的：复用在继承历史之前就已中断，于是 child 为一份自己复制过来、目的恰恰是不必付费的 transcript 付了全额预填充。CLI preset 选择了另一项取舍并保留可继续 fork。想要一个没有继承上下文的长期 child 的部署，本来就有 `spawn`。 **让 `report` 对每个 Agent 可见。** 全局注册会通过让 parent 与 child 拥有相同的 schema 与 section 来恢复逐字节相同的前缀。否决的原因是根 agent、one-shot child、远端 child 与无 agent 调用方都会宣告一件推导不出收件方的工具，而执行期拒绝会让 schema 可见性与权限彼此矛盾——这正是[report 工具 Agent Note](../feature/2026-07-30-continuable-subagent-report-tool.zh.md)已经定下的作用域局部决策。 @@ -38,12 +38,12 @@ one-shot child——前台与后台皆然——经由 `SubagentRuntime.start()` ## 后果 -- 没有任何随附组合会创建可继续的 fork child；`subagent_fork` 把结果返回给调用方的轮次，而 `send_message` 只寻址 spawn 出的 child。 -- 除非部署在 fork 委派工具上配置了 `persona` 或 `toolFilter`，fork child 的请求前缀与其 parent 逐字节相同，因此初始内容的 token 成本重新换来了提供方侧的复用。 -- fork 提供方的可继续路径没有生产调用方，也没有整体组装层面的覆盖。它保留自己的包内测试，seam 也仍然接受它，因此某个组合包或 `--patch` 覆盖层可以无需改动代码、也不会有任何警告地把它重新引入。 +- base 组合包与 ACP/headless 示例只创建 one-shot fork child；其中的 `subagent_fork` 会把结果返回给调用方的轮次，`send_message` 也只寻址 spawn 出的 child。三个 CLI preset 会创建可继续的 fork child。 +- 除非部署在 fork 委派工具上配置了 `persona`、`toolFilter` 或不同的 LLM 路由，one-shot fork child 的请求前缀会与其 parent 逐字节相同，因此初始内容的 token 成本可以换来提供方侧的复用。可继续 fork 会在继承历史之前增加 `report`，从而失去该复用。 +- fork 提供方的可继续路径有 CLI 生产调用方与包内测试。同一条 seam 也接受 one-shot 组合，因此某个组合包或 `--patch` 覆盖层可以无需改动代码、也不会有任何警告地选择任一生命周期。 - `subagent_fork` 面向模型的 schema 发生变化：base 组合包中可继续的后台措辞被 one-shot 的 task 措辞取代，在两个示例中则完全消失。受影响的无密钥快照工具 schema 伴随文件在同一次改动中重新记录。 -- 在随附部署中，report 义务的覆盖范围收窄到 spawn 出的 child。它的 `next-step` 默认调度、权限模型与覆盖仍独立于 fork 组合。 +- 在每个可继续组合中，report 义务都会覆盖 spawn 出的 child；在 CLI preset 中，它也覆盖 fork 出的 child。它的 `next-step` 默认调度、权限模型与覆盖仍独立于 fork 组合。 ### 已接受的风险 -该限制存在于三个配置文件与一处代码注释中，而不在门禁里。未来某个组合包行或 profile 补丁可以在 fork 工具上设置 `backgroundMode: continuable`，从而悄然重新引入前缀损失；没有任何东西会失败得很响亮。这就是不把某一份插件清单的后果写入 `tool-subagent` 所接受的代价。 +one-shot 限制存在于三个配置文件与一处代码注释中，而不在门禁里；CLI preset 行已经选择 `backgroundMode: continuable` 并承担前缀损失。任何组合包或 profile 补丁都能选择任一方式，且不会收到警告。这就是不把某一份插件清单的后果写入 `tool-subagent` 所接受的代价。 diff --git a/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.i18n.yaml b/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.i18n.yaml index 780c4af5ea..57a5f4244b 100644 --- a/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.i18n.yaml +++ b/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.md -2026-06-21-subagent-capability-seam.md: bc84d88d701a5f3018bf00f0ecf8b60750917407 -2026-06-21-subagent-capability-seam.zh.md: 7932181d5667fc8a69ca7aa450fcbf6270ef14d5 +2026-06-21-subagent-capability-seam.md: 70ae99725dffee48292c3481df049563fb54a82c +2026-06-21-subagent-capability-seam.zh.md: 2f63c9022fb116daa2bb6ccd9150d96a823657de diff --git a/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.md b/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.md index bc84d88d70..70ae99725d 100644 --- a/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.md +++ b/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.md @@ -45,7 +45,7 @@ A provider exposes `start(request) → Promise`. Fulfillment publis ### Two kinds of optional capability, discovered two ways -- **Start-time features** (`outputSchema`, `depthLimit`, `toolFilter`, `persona`) ride on a static `provider.capabilities` descriptor. The service checks every requested one BEFORE delegating and **rejects loud** (`SubagentError('UNSUPPORTED_CAPABILITY')`) if the provider lacks it — never accepted-then-ignored. They must be checked before a run exists, which is why they cannot be runtime methods. +- **Start-time features** (`agentOptions`, `outputSchema`, `depthLimit`, `toolFilter`, `persona`) ride on a static `provider.capabilities` descriptor. The service checks every requested one BEFORE delegating and **rejects loud** (`SubagentError('UNSUPPORTED_CAPABILITY')`) if the provider lacks it — never accepted-then-ignored. They must be checked before a run exists, which is why they cannot be runtime methods. - **Continuable creation** is the optional `SubagentProvider.prepareContinuable` method; presence is the capability and TypeScript narrowing is the discovery mechanism, so no separate flag can drift from the implementation. The continuation manager owns later delivery and cold resume directly through `AgentHandle`, while one-shot `SubagentRun` has no steering or resume operation, as refined by [continuable subagents](2026-07-28-continuable-subagent-conversations.md). ### Fork vs. fresh are separate backends, not a flag @@ -60,9 +60,9 @@ Each in-process subagent runs in its **own `Session`** (own id, `parentSession` `dsh-tool-subagent` passes its execution signal to `start()`, awaits the child result, and disposes the run before reporting. Non-completed outcomes become error results rather than successful partial output; they present the optional safe diagnostic owned by the [non-interactive permissions decision](2026-08-15-product-subagent-noninteractive-permissions.md) separately from partial assistant text. Independent result and disposal rejections remain independently observable. -### Provider selection is config, not model-facing +### Transport provider selection is config, not model-facing -`dsh-tool-subagent` binds to exactly one provider name (`Config.provider`); the model sees only `{ description, prompt }`. To expose more than one transport, load the tool plugin more than once, each bound to a different provider and a distinct `toolName` (the tool registry rejects a duplicate name). The *service* holds the multi-provider registry; the *tool* picks one — the schema carries no provider/type parameter. +`dsh-tool-subagent` binds to exactly one subagent transport provider name (`Config.provider`). To expose more than one transport, load the tool plugin more than once, each bound to a different provider and a distinct `toolName` (the tool registry rejects a duplicate name). The *service* holds the multi-provider registry; the *tool* picks one — its schema carries no subagent transport/type parameter. A later opt-in adds child LLM provider/model fields without changing this transport decision; see [model-selected subagent routes](2026-08-18-model-selected-subagent-routes.md). ## Testing diff --git a/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.zh.md b/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.zh.md index 7932181d56..2f63c9022f 100644 --- a/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.zh.md +++ b/.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.zh.md @@ -45,7 +45,7 @@ bash seam（[能力 seam](../architecture/2026-06-13-capability-seams.zh.md)） ### 两类可选能力，两种发现方式 -- **启动时功能**（`outputSchema`、`depthLimit`、`toolFilter`、`persona`）挂在静态的 `provider.capabilities` 描述符上。服务在委派之前检查每个被请求的功能，如果提供方不支持则**响亮拒绝**（`SubagentError('UNSUPPORTED_CAPABILITY')`），绝不接受后静默忽略。这些功能必须在 run 存在之前检查，因此不能是运行时方法。 +- **启动时功能**（`agentOptions`、`outputSchema`、`depthLimit`、`toolFilter`、`persona`）挂在静态的 `provider.capabilities` 描述符上。服务在委派之前检查每个被请求的功能，如果提供方不支持则**响亮拒绝**（`SubagentError('UNSUPPORTED_CAPABILITY')`），绝不接受后静默忽略。这些功能必须在 run 存在之前检查，因此不能是运行时方法。 - **可继续创建**使用可选的 `SubagentProvider.prepareContinuable` 方法；方法是否存在本身即为能力，TypeScript 类型收窄即为发现机制，因此不需要可能与实现失同步的独立 flag。继续执行管理器直接通过 `AgentHandle` 负责后续投递与冷恢复，而一次性 `SubagentRun` 没有 steering 或 resume 操作，具体由[可继续 subagent](2026-07-28-continuable-subagent-conversations.zh.md) 细化。 ### Fork 与 fresh 是独立后端，而非一个 flag @@ -60,9 +60,9 @@ bash seam（[能力 seam](../architecture/2026-06-13-capability-seams.zh.md)） `dsh-tool-subagent` 将其执行信号传给 `start()`，等待子 agent 结果，并在报告前 dispose 该 run。非完成态的结果变为错误结果，而非成功的部分输出；它会把由[非交互权限决策](2026-08-15-product-subagent-noninteractive-permissions.zh.md)负责的可选安全诊断与部分 assistant 文本分开呈现。结果与 dispose 的拒绝仍可彼此独立地观察。 -### 提供方选择是配置，不面向模型 +### 传输提供方选择是配置，不面向模型 -`dsh-tool-subagent` 绑定到恰好一个提供方名称（`Config.provider`）；模型只看到 `{ description, prompt }`。若要暴露多种传输方式，请多次加载该工具插件，每次绑定不同的提供方和不同的 `toolName`（工具注册表拒绝重名）。*服务*持有多提供方注册表；*工具*选择其中一个——schema 中没有提供方/type 参数。 +`dsh-tool-subagent` 绑定到恰好一个 subagent 传输提供方名称（`Config.provider`）。若要暴露多种传输方式，请多次加载该工具插件，每次绑定不同的提供方和不同的 `toolName`（工具注册表拒绝重名）。*服务*持有多提供方注册表；*工具*选择其中一个——schema 中没有 subagent 传输/type 参数。后续 opt-in 增加了子 agent LLM 提供方/模型字段，但没有改变这项传输决策；见[模型选择的 subagent 路由](2026-08-18-model-selected-subagent-routes.zh.md)。 ## 测试 diff --git a/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.i18n.yaml b/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.i18n.yaml new file mode 100644 index 0000000000..d21dd5be39 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.i18n.yaml @@ -0,0 +1,6 @@ +# Bilingual-pair consistency record (docs/i18n/README.md): the git blob hash of each +# side as of the last confirmed-consistent state. Both languages carry equal authority; +# after editing either side, bring the other along and re-record with: +# pnpm run verify-translation-pairing --write .agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md +2026-08-18-model-selected-subagent-routes.md: 1602e3ac90870edbd0206cd87fdf97ecc34cad41 +2026-08-18-model-selected-subagent-routes.zh.md: 0dad8b9d030e6de65cb3fa1e0e93ad7c28bcc5c1 diff --git a/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md b/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md new file mode 100644 index 0000000000..1602e3ac90 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md @@ -0,0 +1,61 @@ +# Agent Note: Model-selected subagent routes + +Status: implemented + +English | [中文](2026-08-18-model-selected-subagent-routes.zh.md) + +## Problem + +`dsh-tool-subagent` can configure child `AgentOptions`, and both in-process providers merge those values over the parent Agent's LLM selection. The model-facing tool could not request a different provider, model, or reasoning effort for one suitable subtask. Loading one distinctly named delegation tool per LLM route duplicates schemas and turns a per-call scheduling choice into deployment configuration. + +The model also needs a bounded way to discover live providers and model-owned effort ids. Rendering the adapter directory into every delegation description would make an advisory, mutable catalog part of the prompt prefix. + +## Decision + +`dsh-tool-subagent` exposes optional `provider`, `model`, and `reasoning_effort` fields only when its instance enables `enableModelSelection`, or its Agent-scoped `modelSelectionSettings` instance resolves an enabled Session decision, and the bound subagent provider advertises `SubagentCapabilities.agentOptions`. No route allowlist is required. Registered LLM provider routes are available for child selection; this tool does not add a second authorization policy over the deployment's LLM registry. Disabled instances omit and reject model-facing selection, while configured `Config.agentOptions` remain deployment-owned defaults. Either selection mode against a provider without the capability fails the plugin mount. + +Provider and model form one route and must be supplied together. An effort may be supplied alone when configured or parent values provide the effective route. Model arguments override `Config.agentOptions`, and configured fields override the parent Agent's latest logged request selection; creation options supply the fallback before its first request and retain the configured output-token limit. Reasoning-effort identifiers remain adapter-owned. An unchanged route inherits an omitted effort, while changing provider or model without naming an effort clears the lower layer's route-owned value so the selected model resolves its own default. `AgentOptions` carries the resulting effort into the child loop, whose request header logs the effective value. A continuable descriptor records it with the resolved provider and model so a child that has not logged its first request can cold-resume with the same selection. + +An explicit or configured provider, model, or effort resolves through `ctx.llm.resolveCallConfig()` before child creation. That lookup owns provider registration, exact-model metadata, reasoning-effort validation, and adapter defaults. The tool checks cancellation again after the asynchronous lookup and before creating a child or background job. Calls with no model-facing selection and no configured route fields preserve the existing provider path without requiring the optional LLM service. + +An enabled definition registers `list_subagent_models`. With no arguments the tool lists registered providers; with `provider` it calls that adapter's advisory model catalog; with `provider` and `model` it resolves the exact model and returns its reasoning efforts and default. At most one instance in a tool scope enables selection because the discovery name is global. Shipped product compositions put `modelSelectionSettings: true` on the primary Agent-scoped `subagent` instance and register the Host-owned `subagent-model-selection` settings namespace with `enabled: false`. A new top-level Session samples that preference during composition and logs an enabled decision as `subagent/model-selection-enabled` before any model request. A child Session inherits the live parent's decision, and a resumed Session uses its existing marker instead of the current preference. Therefore a settings edit affects only subsequently composed top-level Sessions. The fixed discovery definition remains available without the optional LLM service, while discovery and selected-route calls fail until that service is present. An unlisted model remains selectable when the adapter accepts its id. + +Shipped `subagent_fork` instances leave `enableModelSelection` disabled even though the in-process fork provider supports `agentOptions`. A fork inherits the parent's effective provider and model so its copied conversation prefix remains eligible for provider-side KV Cache reuse. Changing either route component requires the new route to prefill that inherited history again, and that recomputation can dominate the delegated task's cost. This restriction is independent of the discovery tool's global name: separating discovery ownership would permit the configuration but would not preserve reuse. Fork route selection remains unavailable until a route change can retain prefix reuse or the caller can explicitly bound and accept the recomputation cost. + +The delegation definition is static across adapter registration and catalog changes, so live topology neither expands every parent request nor invalidates its cache prefix. The discovery result enters the transcript only when called. A custom inheritance-capable instance that enables selection warns that changing provider or model can prevent provider-side reuse of the inherited conversation prefix. + +`SubagentCapabilities.agentOptions` remains the transport truth. The service rejects a request carrying those options before calling a provider that advertises `false`. Both in-process providers advertise `true`; the current ACP, Codex, Claude Code, and DSH SDK transports advertise `false`. Tool configuration that supplies `agentOptions`, statically enables model selection, or makes it settings-controlled also fails when its bound provider lacks the capability. + +## Alternatives considered + +**Keep a deployment-configured route allowlist.** Rejected because it duplicates the live LLM registry, requires configuration before the model can use an already registered route, and creates a second policy surface for clients to edit. Deployments that must restrict LLM access should control which provider routes they register. + +**Render the live adapter catalog in every delegation description.** Rejected because one provider can advertise hundreds of models, inflating every request, and catalog changes would rewrite an early cache-prefix definition. The on-demand directory keeps mutable data out of the fixed schema. + +**Use the advertised catalog as an allowlist.** Rejected because adapter catalogs are advisory and some providers accept arbitrary exact model ids. Exact resolution remains authoritative. + +**Add discovery methods to the subagent service.** Rejected because provider/model/effort metadata already belongs to `ctx.llm`; the new tool is a model-facing consumer of that existing capability. + +**Export discovery as a separately loaded plugin entry.** Rejected because shipped compositions always pair discovery with their primary delegation tool. Explicit ownership on that instance prevents duplicate global names without another Cordis config entry or lifecycle. + +**Configure discovery independently from model-facing selection.** Rejected because discovery exists to supply valid route and effort identifiers to the same model that can select them. One switch prevents a tool schema from advertising selection without its discovery path, or discovery without an applicable delegation route. + +**Enable model-facing route selection on shipped fork tools.** Not shipped because changing provider or model forfeits the inherited prefix's KV Cache reuse and can make prefix recomputation more expensive than the delegated work. The option can be reconsidered when reuse survives the route change or the interface makes that cost explicit and bounded. + +**Use a global reasoning-effort enum.** Rejected because effort identifiers and defaults belong to an exact provider/model route. The LLM adapter validates them without central translation or clamping. + +**Allow remote providers to ignore the fields.** Rejected because the request would claim a route choice that did not happen. The capability flag makes the unsupported path fail before child creation. + +## Consequences + +- An enabled delegation tool can select any live child LLM route without deployment selector configuration; disabled instances omit and reject model-facing route fields. +- The primary delegation-tool instance defaults selection off, exposes a Models-page opt-in for new Sessions, and registers `list_subagent_models` only in Sessions whose durable decision is enabled; its catalog rows do not restrict delegation. +- Shipped fork tools inherit the parent's provider and model and omit model-facing route fields so the inherited conversation prefix remains eligible for KV Cache reuse. +- Omission retains configured defaults and compatible inheritance from the parent's latest logged request; a route change without an explicit effort uses the selected model's default. +- Adapter catalog and topology changes leave the delegation definition and its prompt-cache prefix unchanged. +- Out-of-process subagent providers reject configured and model-selected Agent options until they implement and advertise the capability. +- Unit coverage owns the default-off Host preference, new-Session sampling, child inheritance, resumed decisions, opt-in schema and execution enforcement, merge precedence, route-aware effort inheritance, preflight cancellation, live discovery, diagnostics, definition stability, capability rejection, and optional-service behavior. A shipped headless snapshot pins inheritance from a logged parent selection; the shipped examples also own the assembled keyless model-visible schemas. + +## Related decisions + +This note refines only child LLM routing. The fixed subagent transport remains owned by the [subagent capability seam](2026-06-21-subagent-capability-seam.md), while the separate effect of child-scoped prompt and tool additions on fork prefix reuse remains owned by [cache-preserving forked children stay one-shot](../architecture/2026-08-10-fork-children-stay-one-shot.md). diff --git a/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.zh.md b/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.zh.md new file mode 100644 index 0000000000..0dad8b9d03 --- /dev/null +++ b/.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.zh.md @@ -0,0 +1,61 @@ +# Agent Note: 模型选择的 subagent 路由 + +Status: implemented + +[English](2026-08-18-model-selected-subagent-routes.md) | 中文 + +## 问题 + +`dsh-tool-subagent` 可以配置子级 `AgentOptions`，两个进程内提供方也已把这些值合并到父 Agent 的 LLM 选择之上。但面向模型的工具不能为某个适合的子任务请求不同的提供方、模型或推理强度。为每条 LLM 路由加载一个名称不同的委派工具会重复 schema，并把每次调用的调度选择变成部署配置。 + +模型还需要一种有界方式来发现实时提供方和模型自有的推理强度 ID。把 adapter 目录渲染到每一份委派描述中，会让仅供参考且会变化的目录进入 prompt 前缀。 + +## 决策 + +只有实例启用 `enableModelSelection`，或其 Agent 作用域的 `modelSelectionSettings` 实例解析出已启用的 Session 决定，且绑定的 subagent 提供方声明 `SubagentCapabilities.agentOptions` 时，`dsh-tool-subagent` 才公开可选的 `provider`、`model` 与 `reasoning_effort` 字段，不要求配置路由允许列表。已注册的 LLM 提供方路由都可供子级选择；本工具不会在部署的 LLM 注册表之上增加第二套授权策略。禁用的实例会省略并拒绝面向模型的选择，而配置的 `Config.agentOptions` 仍是部署方所有的默认值。如果提供方缺少该能力，任一种选择模式都会使插件挂载失败。 + +提供方与模型共同组成一条路由，必须一起提供。如果配置值或父级值能够提供生效路由，则可以只提供推理强度。模型参数覆盖 `Config.agentOptions`，配置字段覆盖父 Agent 最新记录的请求选择；首个请求之前由创建选项提供回退，并保留其中配置的输出 token 上限。推理强度 ID 仍由 adapter 所有。路由不变时会继承省略的强度；更换提供方或模型但没有指定强度时，会清除下层路由自有的值，使所选模型解析自己的默认值。`AgentOptions` 把结果强度传入子级循环，其请求 header 会记录生效值。可继续描述符会把它与解析后的提供方和模型一同记录，使尚未写入首个请求的子级能以相同选择冷恢复。 + +显式或配置的提供方、模型或强度会在创建子级前通过 `ctx.llm.resolveCallConfig()` 解析。该查询负责提供方注册、精确模型元数据、推理强度校验和 adapter 默认值。异步查询完成后、创建子级或后台 job 之前，工具会再次检查取消状态。既没有面向模型的选择、也没有配置路由字段的调用会保留原有提供方路径，不要求可选 LLM 服务存在。 + +启用的定义会注册 `list_subagent_models`。无参数调用列出已注册提供方；提供 `provider` 时调用该适配器的建议性模型目录；同时提供 `provider` 与 `model` 时解析精确模型，并返回其推理强度和默认值。因为发现工具使用全局名称，一个工具作用域最多由一个实例启用选择。随附产品组合在 Agent 作用域的主 `subagent` 实例上设置 `modelSelectionSettings: true`，并注册默认 `enabled: false` 的 Host 自有 `subagent-model-selection` settings namespace。新的顶层 Session 会在组合期间读取该偏好，并在任何模型请求之前把启用决定记录为 `subagent/model-selection-enabled`。子 Session 继承在线父级的决定；恢复的 Session 使用已有标记，而不是当前偏好。因此，设置修改只影响之后组合的顶层 Session。即使缺少可选 LLM 服务，固定发现定义仍保持可用；发现调用和所选路由调用会在该服务出现前失败。只要适配器接受某个未列出的模型 ID，仍可选择该模型。 + +随附的 `subagent_fork` 实例不会启用 `enableModelSelection`，即使进程内 fork 提供方支持 `agentOptions` 也是如此。fork 会继承父级生效的提供方与模型，使复制的对话前缀仍可供提供方侧 KV Cache 复用。更改任一路由组件都会要求新路由重新预填充继承的历史，而这项重算成本可能超过委派任务本身。该限制与发现工具的全局名称无关：分离发现工具的持有权可以让配置生效，却无法保留复用。只有在路由变化仍能保留前缀复用，或调用方可以显式限制并接受重算成本时，才重新考虑 fork 路由选择。 + +委派定义不会随 adapter 注册和目录变化而改变，因此实时拓扑既不会扩大每个父级请求，也不会使缓存前缀失效。只有调用发现工具时，目录结果才进入 transcript。自定义的上下文继承实例如果启用选择，其描述会警告，更改提供方或模型可能阻止提供方复用继承的对话前缀。 + +`SubagentCapabilities.agentOptions` 仍是传输事实。如果请求携带这些选项，而提供方声明为 `false`，服务会在调用提供方前拒绝。两个进程内提供方声明为 `true`；当前 ACP、Codex、Claude Code 与 DSH SDK 传输声明为 `false`。工具配置提供 `agentOptions`、静态启用模型选择或让它受 settings 控制时，如果绑定的提供方缺少该能力，也会失败。 + +## 考虑过的替代方案 + +**保留部署配置的路由允许列表。** 不采用，因为它重复实时 LLM 注册表，要求先配置才能让模型使用已经注册的路由，并为客户端增加第二套策略编辑界面。需要限制 LLM 访问的部署应控制所注册的提供方路由。 + +**在每一份委派描述中渲染实时 adapter 目录。** 不采用，因为一个提供方可能公布数百个模型，从而扩大每次请求，而且目录变化会改写缓存前缀中的早期定义。按需目录让可变数据留在固定 schema 之外。 + +**把公布的目录当作允许列表。** 不采用，因为 adapter 目录只提供建议，有些提供方接受任意精确模型 ID。精确解析仍是权威。 + +**在 subagent 服务中增加发现方法。** 不采用，因为提供方／模型／强度元数据已经属于 `ctx.llm`；新工具只是该现有能力面向模型的 Consumer。 + +**把发现工具作为独立加载的插件入口导出。** 不采用，因为随附组合总是把发现工具与主委派工具配套加载。在该实例上显式指定持有权，可以避免重复的全局工具名，无需增加 Cordis 配置项或独立生命周期。 + +**分别配置发现与面向模型的选择。** 不采用，因为发现功能用于向能够选择这些值的同一个模型提供有效的路由与强度 ID。一个开关可以避免工具 schema 公开选择却没有对应发现路径，或公开发现却没有适用的委派路由。 + +**在随附 fork 工具上启用面向模型的路由选择。** 不随产品提供，因为更改提供方或模型会失去继承前缀的 KV Cache 复用，重新预填充前缀的成本可能高于委派工作本身。只有在路由变化仍能保留复用，或接口能把这项成本显式化并限制住时，才重新考虑该选项。 + +**使用全局推理强度枚举。** 不采用，因为推理强度 ID 和默认值属于精确的提供方／模型路由。LLM adapter 会直接校验，无需中心化翻译或截断。 + +**允许远程提供方忽略这些字段。** 不采用，因为请求会声称发生了实际上没有发生的路由选择。能力标记会让不支持的路径在创建子级前失败。 + +## 结果 + +- 启用的委派工具无需部署选择器配置，即可选择任意实时子级 LLM 路由；禁用的实例会省略并拒绝面向模型的路由字段。 +- 主委派工具实例默认关闭选择，为新 Session 提供 Models 页面 opt-in，并且只在持久决定已启用的 Session 中注册 `list_subagent_models`；其目录条目不会限制委派。 +- 随附 fork 工具会继承父级的提供方与模型，并省略面向模型的路由字段，使继承的对话前缀仍可供 KV Cache 复用。 +- 省略选择时保留配置默认值，并从父级最新记录的请求中进行兼容继承；改变路由但不显式指定强度时，使用所选模型的默认值。 +- adapter 目录和拓扑变化不会改变委派定义及其 prompt 缓存前缀。 +- 进程外 subagent 提供方在实现并声明该能力前，会拒绝配置和模型选择的 Agent 选项。 +- 单元测试覆盖默认关闭的 Host 偏好、新 Session 读取、子级继承、恢复决定、选择启用时的 schema 与执行强制、合并优先级、路由相关强度继承、预检取消、实时发现、诊断、定义稳定性、能力拒绝与可选服务行为。随附的 headless 快照固定从父级已记录选择继承的行为；随附示例还覆盖组装后无密钥、模型可见的 schema。 + +## 相关决策 + +本 Note 仅细化子级 LLM 路由。固定的 subagent 传输仍由 [subagent 能力 seam](2026-06-21-subagent-capability-seam.zh.md)负责，而子级作用域提示词与工具增量对 fork 前缀复用产生的独立影响仍由[保留缓存的 fork child 保持 one-shot](../architecture/2026-08-10-fork-children-stay-one-shot.zh.md)负责。 diff --git a/apps/cli/tests/web-agent-presets.e2e.ts b/apps/cli/tests/web-agent-presets.e2e.ts index a4e048b5fd..6af58a3a37 100644 --- a/apps/cli/tests/web-agent-presets.e2e.ts +++ b/apps/cli/tests/web-agent-presets.e2e.ts @@ -11,6 +11,7 @@ import type { Agent } from '@deepseek-ai/dsh-agent' import type { PatchOptions } from '@deepseek-ai/cordis-plugin-include' import { afterAll, beforeAll, describe, expect, it, vi } from 'vitest' import { settingsNamespace } from '@deepseek-ai/dsh-settings' +import { SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE } from '@deepseek-ai/dsh-tool-subagent/model-selection-settings' import { resolveSessionPreset, SETTINGS_NAMESPACE, SHIPPED_PRESET_ROOT } from '@deepseek-ai/dsh-agent-presets' import { applyChildComposition, childSessionMeta } from '@deepseek-ai/dsh-subagent' import { CallId } from '@deepseek-ai/dsh-llm' @@ -240,6 +241,34 @@ describe('the shipped Web composition', () => { } }) + it('applies the default-off subagent model-selection preference only to new sessions', async () => { + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: false }) + const disabled = await ctx.agents.create({ + sessionId: SessionId('preset-model-selection-disabled'), + setup: agentCtx => ctx.agentPresets.mount(agentCtx, 'standard').then(() => undefined), + }) + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: true }) + const enabled = await ctx.agents.create({ + sessionId: SessionId('preset-model-selection-enabled'), + setup: agentCtx => ctx.agentPresets.mount(agentCtx, 'standard').then(() => undefined), + }) + try { + expect(toolNames(ctx, disabled.agent)).not.toContain('list_subagent_models') + expect(toolParameterNames(ctx, disabled.agent, 'subagent')).not.toEqual(expect.arrayContaining([ + 'model', 'provider', 'reasoning_effort', + ])) + expect(toolNames(ctx, enabled.agent)).toContain('list_subagent_models') + expect(toolParameterNames(ctx, enabled.agent, 'subagent')).toEqual(expect.arrayContaining([ + 'model', 'provider', 'reasoning_effort', + ])) + expect(toolNames(ctx, disabled.agent)).not.toContain('list_subagent_models') + } finally { + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: false }) + await enabled.dispose() + await disabled.dispose() + } + }) + it('composes the exact RL prompt and two tools from `minimal`', async () => { const handle = await ctx.agents.create({ sessionId: SessionId('preset-minimal'), diff --git a/apps/web/tests/snapshots/models-settings/configured.expected.md b/apps/web/tests/snapshots/models-settings/configured.expected.md index 3c3be0922c..6ac6d76796 100644 --- a/apps/web/tests/snapshots/models-settings/configured.expected.md +++ b/apps/web/tests/snapshots/models-settings/configured.expected.md @@ -19,6 +19,10 @@ - text: 关闭 - heading "模型" [level=2] - paragraph: 填入各提供方的 API 密钥即可使用其模型。 + - region "Subagent 自选模型": + - heading "Subagent 自选模型" [level=3] + - paragraph: 允许新会话为 subagent 选择提供方、模型和推理强度。运行中的会话不会改变。 + - switch "允许 subagent 自选模型" - status: 已保存 minimax-cn。 - list: - listitem: diff --git a/apps/web/tests/snapshots/models-settings/declared-edit.expected.md b/apps/web/tests/snapshots/models-settings/declared-edit.expected.md index 1d538bcfe4..2f7d4a3a30 100644 --- a/apps/web/tests/snapshots/models-settings/declared-edit.expected.md +++ b/apps/web/tests/snapshots/models-settings/declared-edit.expected.md @@ -19,6 +19,10 @@ - text: 关闭 - heading "模型" [level=2] - paragraph: 填入各提供方的 API 密钥即可使用其模型。 + - region "Subagent 自选模型": + - heading "Subagent 自选模型" [level=3] + - paragraph: 允许新会话为 subagent 选择提供方、模型和推理强度。运行中的会话不会改变。 + - switch "允许 subagent 自选模型" - list: - listitem: - text: minimax-cn diff --git a/apps/web/tests/snapshots/models-settings/declared.expected.md b/apps/web/tests/snapshots/models-settings/declared.expected.md index df48328fd3..bb129bf2ea 100644 --- a/apps/web/tests/snapshots/models-settings/declared.expected.md +++ b/apps/web/tests/snapshots/models-settings/declared.expected.md @@ -19,6 +19,10 @@ - text: 关闭 - heading "模型" [level=2] - paragraph: 填入各提供方的 API 密钥即可使用其模型。 + - region "Subagent 自选模型": + - heading "Subagent 自选模型" [level=3] + - paragraph: 允许新会话为 subagent 选择提供方、模型和推理强度。运行中的会话不会改变。 + - switch "允许 subagent 自选模型" - list: - listitem: - text: minimax-cn diff --git a/apps/web/tests/snapshots/models-settings/empty.expected.md b/apps/web/tests/snapshots/models-settings/empty.expected.md index 54bf1db3c3..dfeb637b5f 100644 --- a/apps/web/tests/snapshots/models-settings/empty.expected.md +++ b/apps/web/tests/snapshots/models-settings/empty.expected.md @@ -19,6 +19,10 @@ - text: 关闭 - heading "模型" [level=2] - paragraph: 填入各提供方的 API 密钥即可使用其模型。 + - region "Subagent 自选模型": + - heading "Subagent 自选模型" [level=3] + - paragraph: 允许新会话为 subagent 选择提供方、模型和推理强度。运行中的会话不会改变。 + - switch "允许 subagent 自选模型" - list - text: 提供方 - combobox "提供方": diff --git a/apps/web/tests/snapshots/onboarding-deepseek-config/models.expected.md b/apps/web/tests/snapshots/onboarding-deepseek-config/models.expected.md index a302932e65..020f70d095 100644 --- a/apps/web/tests/snapshots/onboarding-deepseek-config/models.expected.md +++ b/apps/web/tests/snapshots/onboarding-deepseek-config/models.expected.md @@ -19,6 +19,10 @@ - text: 关闭 - heading "模型" [level=2] - paragraph: 填入各提供方的 API 密钥即可使用其模型。 + - region "Subagent 自选模型": + - heading "Subagent 自选模型" [level=3] + - paragraph: 允许新会话为 subagent 选择提供方、模型和推理强度。运行中的会话不会改变。 + - switch "允许 subagent 自选模型" - list: - listitem: - text: DeepSeek diff --git a/apps/web/tests/snapshots/onboarding-usable-provider/dismissed.expected.md b/apps/web/tests/snapshots/onboarding-usable-provider/dismissed.expected.md index 496443b057..73c66388f3 100644 --- a/apps/web/tests/snapshots/onboarding-usable-provider/dismissed.expected.md +++ b/apps/web/tests/snapshots/onboarding-usable-provider/dismissed.expected.md @@ -19,6 +19,10 @@ - text: 关闭 - heading "模型" [level=2] - paragraph: 填入各提供方的 API 密钥即可使用其模型。 + - region "Subagent 自选模型": + - heading "Subagent 自选模型" [level=3] + - paragraph: 允许新会话为 subagent 选择提供方、模型和推理强度。运行中的会话不会改变。 + - switch "允许 subagent 自选模型" - list: - listitem: - text: DeepSeek diff --git a/docs/capability-seams.i18n.yaml b/docs/capability-seams.i18n.yaml index 406b4015a0..d7e1c70e5a 100644 --- a/docs/capability-seams.i18n.yaml +++ b/docs/capability-seams.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/capability-seams.md -capability-seams.md: 75f050f329e709e5c88bffbe0d3bc2072d4286de -capability-seams.zh.md: 25fa48c67e406b03677debba44eff5d49fd3c626 +capability-seams.md: 0994b186f7daa6afbff0f1484216e55fc79170ff +capability-seams.zh.md: 48aa0a5353bba8d56f63abfcfc4cd2c601569cb6 diff --git a/docs/capability-seams.md b/docs/capability-seams.md index 75f050f329..0994b186f7 100644 --- a/docs/capability-seams.md +++ b/docs/capability-seams.md @@ -56,6 +56,8 @@ flowchart LR pkg_settings["settings"] svc_settings["ctx.settings
User-settings seam"] pkg_settings_file["settings-file"] + pkg_tool_subagent["tool-subagent"] + svc_subagentModelSelection["ctx.subagentModelSelection
Subagent model-selection preference"] pkg_credentials["credentials"] svc_credentials["ctx.credentials
Credential seam"] pkg_credentials_local["credentials-local"] @@ -94,7 +96,6 @@ flowchart LR pkg_tool_ask_user["tool-ask-user"] pkg_tool_cordis["tool-cordis"] pkg_tool_skill["tool-skill"] - pkg_tool_subagent["tool-subagent"] pkg_tool_todo["tool-todo"] pkg_user_questions["user-questions"] svc_userQuestions["ctx.userQuestions
Human question/answer seam"] @@ -310,6 +311,7 @@ flowchart LR pkg_terminal --> svc_terminals pkg_terminal_bash --> svc_terminals pkg_token_meter --> svc_tokenMeter + pkg_tool_subagent --> svc_subagentModelSelection pkg_tools --> svc_tools pkg_typert_registry --> svc_typert pkg_user_questions --> svc_userQuestions @@ -401,6 +403,7 @@ flowchart LR svc_storage --> pkg_storage_domain svc_storageDomain --> pkg_message_feedback svc_storageDomain --> pkg_workspace + svc_subagentModelSelection --> pkg_tool_subagent svc_subagents --> pkg_tool_ralph svc_subagents --> pkg_tool_subagent svc_subagents --> pkg_tool_subagent_control @@ -458,6 +461,7 @@ flowchart LR | `ctx.typertGateway` | `core` | [`api-gateway`](../packages/api/gateway) | - | - | - | Associates generated Remote descriptors with live Cordis services, resolves registered identities, and exposes unary calls through the shared Connection RPC carrier. | | `ctx.sessionPersistence` | `seam` | [`session-persistence`](../packages/session/session-persistence) | [`session-persistence-jsonl`](../packages/session/session-persistence-jsonl), [`session-persistence-sqlite`](../packages/session/session-persistence-sqlite) | [`agent-loop`](../packages/core/agent-loop), [`tool-bash`](../packages/shell/tool-bash), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex), [`session-query`](../packages/session-query/session-query), [`session-query-sqlite`](../packages/session-query/session-query-sqlite), [`message-feedback`](../packages/feedback/message-feedback) | - | Backends persist the same SessionEvent vocabulary; apps choose a backend at composition time. | | `ctx.settings` | `seam` | [`settings`](../packages/settings/settings) | [`settings-file`](../packages/settings/settings-file) | [`llm-deepseek`](../packages/llm/llm-deepseek), [`llm-pi-ai`](../packages/llm/llm-pi-ai), `apiproxy` | - | Plugins register namespace schemas and resolve layered values; providers store the raw document. The LLM adapters register their entry config as the composition base under the user section; the web gateway serves redacted layered descriptors and writes the user layer. | +| `ctx.subagentModelSelection` | `core` | [`tool-subagent`](../packages/subagent/tool-subagent) | - | [`tool-subagent`](../packages/subagent/tool-subagent) | - | Owns the default-off settings namespace that Agent-scoped delegation tools sample when composing a new top-level Session. | | `ctx.credentials` | `seam` | [`credentials`](../packages/credentials/credentials) | [`credentials-local`](../packages/credentials/credentials-local) | [`llm-deepseek`](../packages/llm/llm-deepseek), [`llm-pi-ai`](../packages/llm/llm-pi-ai), `apiproxy` | - | Configuration carries references to secrets; providers own the values. Consumers resolve per operation, so a rotated credential reaches the very next request; the web gateway exposes value-free views and write-only storage. | | `ctx.authorization` | `seam` | [`authorization`](../packages/credentials/authorization) | - | [`llm-pi-ai`](../packages/llm/llm-pi-ai) | - | Flows are registered by the plugin that knows how to obtain one credential and keyed by the record they write; the seam owns the conversation and the one-attempt-per-key lifecycle, never the protocol. | | `ctx.sessionTelemetry` | `seam` | [`session-telemetry`](../packages/session/session-telemetry) | [`session-telemetry-otel`](../packages/session/session-telemetry-otel) | - | - | The seam captures, redacts, and hands session records to one backend; nothing else consumes the service — its output leaves the process. | diff --git a/docs/capability-seams.zh.md b/docs/capability-seams.zh.md index 25fa48c67e..48aa0a5353 100644 --- a/docs/capability-seams.zh.md +++ b/docs/capability-seams.zh.md @@ -58,6 +58,8 @@ flowchart LR pkg_settings["settings"] svc_settings["ctx.settings
User-settings seam"] pkg_settings_file["settings-file"] + pkg_tool_subagent["tool-subagent"] + svc_subagentModelSelection["ctx.subagentModelSelection
Subagent model-selection preference"] pkg_credentials["credentials"] svc_credentials["ctx.credentials
Credential seam"] pkg_credentials_local["credentials-local"] @@ -96,7 +98,6 @@ flowchart LR pkg_tool_ask_user["tool-ask-user"] pkg_tool_cordis["tool-cordis"] pkg_tool_skill["tool-skill"] - pkg_tool_subagent["tool-subagent"] pkg_tool_todo["tool-todo"] pkg_user_questions["user-questions"] svc_userQuestions["ctx.userQuestions
Human question/answer seam"] @@ -312,6 +313,7 @@ flowchart LR pkg_terminal --> svc_terminals pkg_terminal_bash --> svc_terminals pkg_token_meter --> svc_tokenMeter + pkg_tool_subagent --> svc_subagentModelSelection pkg_tools --> svc_tools pkg_typert_registry --> svc_typert pkg_user_questions --> svc_userQuestions @@ -403,6 +405,7 @@ flowchart LR svc_storage --> pkg_storage_domain svc_storageDomain --> pkg_message_feedback svc_storageDomain --> pkg_workspace + svc_subagentModelSelection --> pkg_tool_subagent svc_subagents --> pkg_tool_ralph svc_subagents --> pkg_tool_subagent svc_subagents --> pkg_tool_subagent_control @@ -460,6 +463,7 @@ flowchart LR | `ctx.typertGateway` | `core` | [`api-gateway`](../packages/api/gateway) | - | - | - | 将生成的 Remote 描述符与实时 Cordis 服务关联，解析已注册的身份，并通过共享的 Connection RPC 载体提供一元调用。 | | `ctx.sessionPersistence` | `seam` | [`session-persistence`](../packages/session/session-persistence) | [`session-persistence-jsonl`](../packages/session/session-persistence-jsonl), [`session-persistence-sqlite`](../packages/session/session-persistence-sqlite) | [`agent-loop`](../packages/core/agent-loop), [`tool-bash`](../packages/shell/tool-bash), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex), [`session-query`](../packages/session-query/session-query), [`session-query-sqlite`](../packages/session-query/session-query-sqlite), [`message-feedback`](../packages/feedback/message-feedback) | - | 各后端持久化同一套 SessionEvent 词汇；应用在组合时选择后端。 | | `ctx.settings` | `seam` | [`settings`](../packages/settings/settings) | [`settings-file`](../packages/settings/settings-file) | [`llm-deepseek`](../packages/llm/llm-deepseek), [`llm-pi-ai`](../packages/llm/llm-pi-ai), `apiproxy` | - | 插件注册命名空间 schema 并解析分层值；提供方存储原始文档。LLM（大语言模型）适配器在用户分区下将其入口配置注册为组合基础；Web 网关提供经过脱敏的分层描述符，并写入用户层。 | +| `ctx.subagentModelSelection` | `core` | [`tool-subagent`](../packages/subagent/tool-subagent) | - | [`tool-subagent`](../packages/subagent/tool-subagent) | - | 拥有默认关闭的设置命名空间；Agent 作用域的委派工具会在组合新顶层 Session 时读取它。 | | `ctx.credentials` | `seam` | [`credentials`](../packages/credentials/credentials) | [`credentials-local`](../packages/credentials/credentials-local) | [`llm-deepseek`](../packages/llm/llm-deepseek), [`llm-pi-ai`](../packages/llm/llm-pi-ai), `apiproxy` | - | 配置携带对机密信息的引用；提供方拥有实际值。消费方按操作解析，因此轮换后的凭据会在紧接着的下一次请求中生效；Web 网关提供不含实际值的视图和只写存储。 | | `ctx.authorization` | `seam` | [`authorization`](../packages/credentials/authorization) | - | [`llm-pi-ai`](../packages/llm/llm-pi-ai) | - | flow 由知道如何取得某份凭据的插件注册，并以其写入的记录为键；seam 拥有这段对话与"每个键同时只跑一次尝试"的生命周期，而非协议本身。 | | `ctx.sessionTelemetry` | `seam` | [`session-telemetry`](../packages/session/session-telemetry) | [`session-telemetry-otel`](../packages/session/session-telemetry-otel) | - | - | 该 seam 捕获会话记录、进行脱敏并交给一个后端；没有其他组件消费该服务，其输出会离开当前进程。 | diff --git a/docs/config-catalog.i18n.yaml b/docs/config-catalog.i18n.yaml index eabedcfe3c..9ba399370d 100644 --- a/docs/config-catalog.i18n.yaml +++ b/docs/config-catalog.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/config-catalog.md -config-catalog.md: 7f7e5d3c58953ea50c43eed0d903f0d5469e7849 -config-catalog.zh.md: f9ab7a8537e29cb0e74e05e74b4a7890d146c28b +config-catalog.md: 6494070f6204ce6130e22d6f71ddbfd18fdc69dd +config-catalog.zh.md: da19d854ca520425a799cb699238a7dca05a90da diff --git a/docs/config-catalog.md b/docs/config-catalog.md index 7f7e5d3c58..6494070f62 100644 --- a/docs/config-catalog.md +++ b/docs/config-catalog.md @@ -958,7 +958,7 @@ export interface DeepSeekCatalogModel { Depends on: [`ModelModality`](../packages/llm/llm/src/index.ts) · [`RetryPolicyConfig`](../packages/llm/llm/src/index.ts) -Source: [`packages/llm/llm-deepseek/src/index.ts:107`](../packages/llm/llm-deepseek/src/index.ts) +Source: [`packages/llm/llm-deepseek/src/index.ts:117`](../packages/llm/llm-deepseek/src/index.ts) @@ -2841,6 +2841,14 @@ export interface Config { * a distinct name. */ toolName?: string + /** Let the model discover and select the child LLM route (default false). */ + enableModelSelection?: boolean + /** + * Sample the Host `subagent-model-selection` user setting for each new + * top-level session and inherit that decision in its child sessions. Mutually + * exclusive with `enableModelSelection`. + */ + modelSelectionSettings?: boolean /** * Expose `run_in_background` (default true). Disabled instances omit the * parameter and reject forced background calls. @@ -2888,7 +2896,7 @@ export interface Config { Depends on: [`AgentOptions`](subsystems/core.md) -Source: [`packages/subagent/tool-subagent/src/index.ts:29`](../packages/subagent/tool-subagent/src/index.ts) +Source: [`packages/subagent/tool-subagent/src/index.ts:48`](../packages/subagent/tool-subagent/src/index.ts) diff --git a/docs/config-catalog.zh.md b/docs/config-catalog.zh.md index f9ab7a8537..da19d854ca 100644 --- a/docs/config-catalog.zh.md +++ b/docs/config-catalog.zh.md @@ -2843,6 +2843,14 @@ export interface Config { * a distinct name. */ toolName?: string + /** Let the model discover and select the child LLM route (default false). */ + enableModelSelection?: boolean + /** + * Sample the Host `subagent-model-selection` user setting for each new + * top-level session and inherit that decision in its child sessions. Mutually + * exclusive with `enableModelSelection`. + */ + modelSelectionSettings?: boolean /** * Expose `run_in_background` (default true). Disabled instances omit the * parameter and reject forced background calls. @@ -2890,7 +2898,7 @@ export interface Config { 依赖：[`AgentOptions`](subsystems/core.zh.md) -来源：[`packages/subagent/tool-subagent/src/index.ts:29`](../packages/subagent/tool-subagent/src/index.ts) +来源：[`packages/subagent/tool-subagent/src/index.ts:48`](../packages/subagent/tool-subagent/src/index.ts) diff --git a/docs/event-producer-consumer.i18n.yaml b/docs/event-producer-consumer.i18n.yaml index c9e637910e..33c536f50a 100644 --- a/docs/event-producer-consumer.i18n.yaml +++ b/docs/event-producer-consumer.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/event-producer-consumer.md -event-producer-consumer.md: 2be5a84969b9f14823abf90cf289a0a41e48dd11 -event-producer-consumer.zh.md: 5bbae1be5d03c3e443d36093ce60dbf7e4b07971 +event-producer-consumer.md: 4ab1275e9309e150504d6a3bdd80792bb1a49ea1 +event-producer-consumer.zh.md: 8d8a401bfdccc74d5774ce5ce7ceac2c548f6c72 diff --git a/docs/event-producer-consumer.md b/docs/event-producer-consumer.md index 2be5a84969..4ab1275e93 100644 --- a/docs/event-producer-consumer.md +++ b/docs/event-producer-consumer.md @@ -9,18 +9,18 @@ This matrix shows which packages dispatch each harness-owned event and which pac | --- | --- | --- | --- | --- | | `agent-loop/config-start-failed` | `emit` | [`packages/core/agent-loop/src/index.ts:183`](../packages/core/agent-loop/src/index.ts) | [`agent-loop`](../packages/core/agent-loop) (`events.dispatch`) | - | | `agent-preset/selected` | `emit` | [`packages/preset/agent-presets/src/types.ts:13`](../packages/preset/agent-presets/src/types.ts) | [`agent-presets`](../packages/preset/agent-presets) (`emit`) | `remotes` | -| `agent/created` | `emit` | [`packages/core/agent/src/runtime-types.ts:159`](../packages/core/agent/src/runtime-types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-presets`](../packages/preset/agent-presets), [`file-reference-local`](../packages/context/file-reference-local), [`goal-round-driver`](../packages/goal/goal-round-driver), [`schedule`](../packages/schedule/schedule), `tool-agent-team` | -| `agent/disposed` | `emit` | [`packages/core/agent/src/runtime-types.ts:168`](../packages/core/agent/src/runtime-types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-loop`](../packages/core/agent-loop), [`file-reference-local`](../packages/context/file-reference-local), [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent), `tool-agent-team` | -| `agent/error` | `emit` | [`packages/core/agent/src/runtime-types.ts:290`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`acp`](../packages/acp/acp), [`goal-round-driver`](../packages/goal/goal-round-driver), `session-controller`, [`session-telemetry`](../packages/session/session-telemetry) | -| `agent/inbox/claimed` | `emit` | [`packages/core/agent/src/runtime-types.ts:197`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`acp`](../packages/acp/acp), [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent), [`tool-jobs`](../packages/jobs/tool-jobs) | -| `agent/inbox/discarded` | `emit` | [`packages/core/agent/src/runtime-types.ts:205`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent) | -| `agent/inbox/inserted` | `emit` | [`packages/core/agent/src/runtime-types.ts:186`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`goal-round-driver`](../packages/goal/goal-round-driver) | -| `agent/pre-step` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:231`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent-instructions`](../packages/context/agent-instructions), [`compaction-basic`](../packages/compaction/compaction-basic), [`goal-round-driver`](../packages/goal/goal-round-driver), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex), [`plan-mode`](../packages/plan/plan-mode), [`repeat-tool-reminder`](../packages/guard/repeat-tool-reminder), [`session-checkpoint-policy`](../packages/session/session-checkpoint-policy), [`session-reference`](../packages/context/session-reference), [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver), [`time-context`](../packages/context/time-context), [`tmux-context`](../packages/context/tmux-context), [`tool-cordis`](../packages/extensions/tool-cordis), [`tool-skill`](../packages/skill/tool-skill) | -| `agent/request` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:244`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent`](../packages/core/agent), [`webhook`](../packages/webhook/webhook) | -| `agent/request-error` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:260`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`compaction-basic`](../packages/compaction/compaction-basic), [`llm-retry`](../packages/llm/llm-retry) | -| `agent/session-start` | `emit` | [`packages/core/agent/src/runtime-types.ts:217`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | `agent-team`, [`goal`](../packages/goal/goal), [`goal-round-driver`](../packages/goal/goal-round-driver), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex) | -| `agent/status` | `emit` | [`packages/core/agent/src/runtime-types.ts:178`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`agent`](../packages/core/agent), `agent-team`, [`compaction-basic`](../packages/compaction/compaction-basic), [`goal-round-driver`](../packages/goal/goal-round-driver), [`schedule`](../packages/schedule/schedule), `server`, `session-controller` | -| `agent/turn-stopping` | `serial` | [`packages/core/agent/src/runtime-types.ts:278`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`serial`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex) | +| `agent/created` | `emit` | [`packages/core/agent/src/runtime-types.ts:161`](../packages/core/agent/src/runtime-types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-presets`](../packages/preset/agent-presets), [`file-reference-local`](../packages/context/file-reference-local), [`goal-round-driver`](../packages/goal/goal-round-driver), [`schedule`](../packages/schedule/schedule), `tool-agent-team`, [`tool-subagent`](../packages/subagent/tool-subagent) | +| `agent/disposed` | `emit` | [`packages/core/agent/src/runtime-types.ts:170`](../packages/core/agent/src/runtime-types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-loop`](../packages/core/agent-loop), [`file-reference-local`](../packages/context/file-reference-local), [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent), `tool-agent-team`, [`tool-subagent`](../packages/subagent/tool-subagent) | +| `agent/error` | `emit` | [`packages/core/agent/src/runtime-types.ts:292`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`acp`](../packages/acp/acp), [`goal-round-driver`](../packages/goal/goal-round-driver), `session-controller`, [`session-telemetry`](../packages/session/session-telemetry) | +| `agent/inbox/claimed` | `emit` | [`packages/core/agent/src/runtime-types.ts:199`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`acp`](../packages/acp/acp), [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent), [`tool-jobs`](../packages/jobs/tool-jobs) | +| `agent/inbox/discarded` | `emit` | [`packages/core/agent/src/runtime-types.ts:207`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent) | +| `agent/inbox/inserted` | `emit` | [`packages/core/agent/src/runtime-types.ts:188`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`goal-round-driver`](../packages/goal/goal-round-driver) | +| `agent/pre-step` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:233`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent-instructions`](../packages/context/agent-instructions), [`compaction-basic`](../packages/compaction/compaction-basic), [`goal-round-driver`](../packages/goal/goal-round-driver), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex), [`plan-mode`](../packages/plan/plan-mode), [`repeat-tool-reminder`](../packages/guard/repeat-tool-reminder), [`session-checkpoint-policy`](../packages/session/session-checkpoint-policy), [`session-reference`](../packages/context/session-reference), [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver), [`time-context`](../packages/context/time-context), [`tmux-context`](../packages/context/tmux-context), [`tool-cordis`](../packages/extensions/tool-cordis), [`tool-skill`](../packages/skill/tool-skill), [`tool-subagent`](../packages/subagent/tool-subagent) | +| `agent/request` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:246`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent`](../packages/core/agent), [`webhook`](../packages/webhook/webhook) | +| `agent/request-error` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:262`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`compaction-basic`](../packages/compaction/compaction-basic), [`llm-retry`](../packages/llm/llm-retry) | +| `agent/session-start` | `emit` | [`packages/core/agent/src/runtime-types.ts:219`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | `agent-team`, [`goal`](../packages/goal/goal), [`goal-round-driver`](../packages/goal/goal-round-driver), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex) | +| `agent/status` | `emit` | [`packages/core/agent/src/runtime-types.ts:180`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`agent`](../packages/core/agent), `agent-team`, [`compaction-basic`](../packages/compaction/compaction-basic), [`goal-round-driver`](../packages/goal/goal-round-driver), [`schedule`](../packages/schedule/schedule), `server`, `session-controller` | +| `agent/turn-stopping` | `serial` | [`packages/core/agent/src/runtime-types.ts:280`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`serial`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex) | | `api-session/activity` | `emit` | [`packages/api/session-controller/src/types.ts:444`](../packages/api/session-controller/src/types.ts) | `session-controller` (`emit`) | `remotes` | | `api-session/added` | `emit` | [`packages/api/session-controller/src/types.ts:424`](../packages/api/session-controller/src/types.ts) | `session-controller` (`emit`) | `remotes` | | `api-session/error` | `emit` | [`packages/api/session-controller/src/types.ts:451`](../packages/api/session-controller/src/types.ts) | `session-controller` (`emit`) | `remotes` | @@ -52,13 +52,13 @@ This matrix shows which packages dispatch each harness-owned event and which pac | `settings/document-updated` | `emit` | [`packages/settings/settings/src/types.ts:48`](../packages/settings/settings/src/types.ts) | [`settings`](../packages/settings/settings) (`events.dispatch`) | `remotes` | | `settings/updated` | `emit` | [`packages/settings/settings/src/types.ts:35`](../packages/settings/settings/src/types.ts) | [`settings`](../packages/settings/settings) (`events.dispatch`) | [`settings`](../packages/settings/settings) | | `skills/change` | `emit` | [`packages/skill/skill/src/index.ts:297`](../packages/skill/skill/src/index.ts) | [`skill`](../packages/skill/skill) (`events.dispatch`) | - | -| `subagent/end` | `emit` | [`packages/subagent/subagent/src/index.ts:164`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), `server`, [`subagent`](../packages/subagent/subagent) | -| `subagent/provider-added` | `emit` | [`packages/subagent/subagent/src/index.ts:138`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`emit`) | [`subagent`](../packages/subagent/subagent), [`tool-subagent`](../packages/subagent/tool-subagent) | -| `subagent/provider-removed` | `emit` | [`packages/subagent/subagent/src/index.ts:144`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`subagent`](../packages/subagent/subagent), [`tool-subagent`](../packages/subagent/tool-subagent) | -| `subagent/start` | `emit` | [`packages/subagent/subagent/src/index.ts:155`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`subagent`](../packages/subagent/subagent) | +| `subagent/end` | `emit` | [`packages/subagent/subagent/src/index.ts:165`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), `server`, [`subagent`](../packages/subagent/subagent) | +| `subagent/provider-added` | `emit` | [`packages/subagent/subagent/src/index.ts:139`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`emit`) | [`subagent`](../packages/subagent/subagent), [`tool-subagent`](../packages/subagent/tool-subagent) | +| `subagent/provider-removed` | `emit` | [`packages/subagent/subagent/src/index.ts:145`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`subagent`](../packages/subagent/subagent), [`tool-subagent`](../packages/subagent/tool-subagent) | +| `subagent/start` | `emit` | [`packages/subagent/subagent/src/index.ts:156`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`subagent`](../packages/subagent/subagent) | | `system-prompt/assemble` | `waterfall` | [`packages/core/system-prompt/src/index.ts:31`](../packages/core/system-prompt/src/index.ts) | [`system-prompt`](../packages/core/system-prompt) (`waterfall`) | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`system-prompt`](../packages/core/system-prompt) | | `system-prompt/change` | `emit` | [`packages/core/system-prompt/src/index.ts:37`](../packages/core/system-prompt/src/index.ts) | [`system-prompt`](../packages/core/system-prompt) (`emit`) | - | -| `tools/change` | `emit` | [`packages/core/tools/src/index.ts:207`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`emit`) | - | +| `tools/change` | `emit` | [`packages/core/tools/src/index.ts:207`](../packages/core/tools/src/index.ts) | [`agent-presets`](../packages/preset/agent-presets) (`emit`), [`tools`](../packages/core/tools) (`emit`) | [`tool-subagent`](../packages/subagent/tool-subagent) | | `tools/code-dispatch-log` | `waterfall` | [`packages/core/tools/src/index.ts:189`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`waterfall`) | [`spill-policy`](../packages/spill/spill-policy) | | `tools/execute` | `waterfall` | [`packages/core/tools/src/index.ts:163`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`waterfall`) | [`session-checkpoint-policy`](../packages/session/session-checkpoint-policy), `timeout-policy` | | `tools/post-execute` | `waterfall` | [`packages/core/tools/src/index.ts:175`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`waterfall`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex), [`repeat-tool-reminder`](../packages/guard/repeat-tool-reminder), [`spill-policy`](../packages/spill/spill-policy), [`tool-fs-search`](../packages/fs/tool-fs-search) | diff --git a/docs/event-producer-consumer.zh.md b/docs/event-producer-consumer.zh.md index 5bbae1be5d..8d8a401bfd 100644 --- a/docs/event-producer-consumer.zh.md +++ b/docs/event-producer-consumer.zh.md @@ -11,13 +11,13 @@ | --- | --- | --- | --- | --- | | `agent-loop/config-start-failed` | `emit` | [`packages/core/agent-loop/src/index.ts:183`](../packages/core/agent-loop/src/index.ts) | [`agent-loop`](../packages/core/agent-loop) (`events.dispatch`) | - | | `agent-preset/selected` | `emit` | [`packages/preset/agent-presets/src/types.ts:13`](../packages/preset/agent-presets/src/types.ts) | [`agent-presets`](../packages/preset/agent-presets) (`emit`) | `remotes` | -| `agent/created` | `emit` | [`packages/core/agent/src/runtime-types.ts:159`](../packages/core/agent/src/runtime-types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-presets`](../packages/preset/agent-presets), [`file-reference-local`](../packages/context/file-reference-local), [`goal-round-driver`](../packages/goal/goal-round-driver), [`schedule`](../packages/schedule/schedule), `tool-agent-team` | -| `agent/disposed` | `emit` | [`packages/core/agent/src/runtime-types.ts:168`](../packages/core/agent/src/runtime-types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-loop`](../packages/core/agent-loop), [`file-reference-local`](../packages/context/file-reference-local), [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent), `tool-agent-team` | +| `agent/created` | `emit` | [`packages/core/agent/src/runtime-types.ts:161`](../packages/core/agent/src/runtime-types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-presets`](../packages/preset/agent-presets), [`file-reference-local`](../packages/context/file-reference-local), [`goal-round-driver`](../packages/goal/goal-round-driver), [`schedule`](../packages/schedule/schedule), `tool-agent-team`, [`tool-subagent`](../packages/subagent/tool-subagent) | +| `agent/disposed` | `emit` | [`packages/core/agent/src/runtime-types.ts:170`](../packages/core/agent/src/runtime-types.ts) | [`agent`](../packages/core/agent) (`events.dispatch`) | [`agent-loop`](../packages/core/agent-loop), [`file-reference-local`](../packages/context/file-reference-local), [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent), `tool-agent-team`, [`tool-subagent`](../packages/subagent/tool-subagent) | | `agent/error` | `emit` | [`packages/core/agent/src/runtime-types.ts:290`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`acp`](../packages/acp/acp), [`goal-round-driver`](../packages/goal/goal-round-driver), `session-controller`, [`session-telemetry`](../packages/session/session-telemetry) | | `agent/inbox/claimed` | `emit` | [`packages/core/agent/src/runtime-types.ts:197`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`acp`](../packages/acp/acp), [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent), [`tool-jobs`](../packages/jobs/tool-jobs) | | `agent/inbox/discarded` | `emit` | [`packages/core/agent/src/runtime-types.ts:205`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`goal-round-driver`](../packages/goal/goal-round-driver), [`subagent`](../packages/subagent/subagent) | | `agent/inbox/inserted` | `emit` | [`packages/core/agent/src/runtime-types.ts:186`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emit`) | [`goal-round-driver`](../packages/goal/goal-round-driver) | -| `agent/pre-step` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:231`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent-instructions`](../packages/context/agent-instructions), [`compaction-basic`](../packages/compaction/compaction-basic), [`goal-round-driver`](../packages/goal/goal-round-driver), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex), [`plan-mode`](../packages/plan/plan-mode), [`repeat-tool-reminder`](../packages/guard/repeat-tool-reminder), [`session-checkpoint-policy`](../packages/session/session-checkpoint-policy), [`session-reference`](../packages/context/session-reference), [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver), [`time-context`](../packages/context/time-context), [`tmux-context`](../packages/context/tmux-context), [`tool-cordis`](../packages/extensions/tool-cordis), [`tool-skill`](../packages/skill/tool-skill) | +| `agent/pre-step` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:231`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent-instructions`](../packages/context/agent-instructions), [`compaction-basic`](../packages/compaction/compaction-basic), [`goal-round-driver`](../packages/goal/goal-round-driver), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex), [`plan-mode`](../packages/plan/plan-mode), [`repeat-tool-reminder`](../packages/guard/repeat-tool-reminder), [`session-checkpoint-policy`](../packages/session/session-checkpoint-policy), [`session-reference`](../packages/context/session-reference), [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver), [`time-context`](../packages/context/time-context), [`tmux-context`](../packages/context/tmux-context), [`tool-cordis`](../packages/extensions/tool-cordis), [`tool-skill`](../packages/skill/tool-skill), [`tool-subagent`](../packages/subagent/tool-subagent) | | `agent/request` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:244`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`agent`](../packages/core/agent), [`webhook`](../packages/webhook/webhook) | | `agent/request-error` | `waterfall` | [`packages/core/agent/src/runtime-types.ts:260`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`waterfall`) | [`compaction-basic`](../packages/compaction/compaction-basic), [`llm-retry`](../packages/llm/llm-retry) | | `agent/session-start` | `emit` | [`packages/core/agent/src/runtime-types.ts:217`](../packages/core/agent/src/runtime-types.ts) | [`agent-loop`](../packages/core/agent-loop) (`emitAgentEvent`) | `agent-team`, [`goal`](../packages/goal/goal), [`goal-round-driver`](../packages/goal/goal-round-driver), [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex) | @@ -54,13 +54,13 @@ | `settings/document-updated` | `emit` | [`packages/settings/settings/src/types.ts:48`](../packages/settings/settings/src/types.ts) | [`settings`](../packages/settings/settings) (`events.dispatch`) | `remotes` | | `settings/updated` | `emit` | [`packages/settings/settings/src/types.ts:35`](../packages/settings/settings/src/types.ts) | [`settings`](../packages/settings/settings) (`events.dispatch`) | [`settings`](../packages/settings/settings) | | `skills/change` | `emit` | [`packages/skill/skill/src/index.ts:297`](../packages/skill/skill/src/index.ts) | [`skill`](../packages/skill/skill) (`events.dispatch`) | - | -| `subagent/end` | `emit` | [`packages/subagent/subagent/src/index.ts:164`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), `server`, [`subagent`](../packages/subagent/subagent) | -| `subagent/provider-added` | `emit` | [`packages/subagent/subagent/src/index.ts:138`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`emit`) | [`subagent`](../packages/subagent/subagent), [`tool-subagent`](../packages/subagent/tool-subagent) | -| `subagent/provider-removed` | `emit` | [`packages/subagent/subagent/src/index.ts:144`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`subagent`](../packages/subagent/subagent), [`tool-subagent`](../packages/subagent/tool-subagent) | -| `subagent/start` | `emit` | [`packages/subagent/subagent/src/index.ts:155`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`subagent`](../packages/subagent/subagent) | +| `subagent/end` | `emit` | [`packages/subagent/subagent/src/index.ts:165`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), `server`, [`subagent`](../packages/subagent/subagent) | +| `subagent/provider-added` | `emit` | [`packages/subagent/subagent/src/index.ts:139`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`emit`) | [`subagent`](../packages/subagent/subagent), [`tool-subagent`](../packages/subagent/tool-subagent) | +| `subagent/provider-removed` | `emit` | [`packages/subagent/subagent/src/index.ts:145`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`subagent`](../packages/subagent/subagent), [`tool-subagent`](../packages/subagent/tool-subagent) | +| `subagent/start` | `emit` | [`packages/subagent/subagent/src/index.ts:156`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`subagent`](../packages/subagent/subagent) | | `system-prompt/assemble` | `waterfall` | [`packages/core/system-prompt/src/index.ts:31`](../packages/core/system-prompt/src/index.ts) | [`system-prompt`](../packages/core/system-prompt) (`waterfall`) | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`system-prompt`](../packages/core/system-prompt) | | `system-prompt/change` | `emit` | [`packages/core/system-prompt/src/index.ts:37`](../packages/core/system-prompt/src/index.ts) | [`system-prompt`](../packages/core/system-prompt) (`emit`) | - | -| `tools/change` | `emit` | [`packages/core/tools/src/index.ts:207`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`emit`) | - | +| `tools/change` | `emit` | [`packages/core/tools/src/index.ts:207`](../packages/core/tools/src/index.ts) | [`agent-presets`](../packages/preset/agent-presets) (`emit`), [`tools`](../packages/core/tools) (`emit`) | [`tool-subagent`](../packages/subagent/tool-subagent) | | `tools/code-dispatch-log` | `waterfall` | [`packages/core/tools/src/index.ts:189`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`waterfall`) | [`spill-policy`](../packages/spill/spill-policy) | | `tools/execute` | `waterfall` | [`packages/core/tools/src/index.ts:163`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`waterfall`) | [`session-checkpoint-policy`](../packages/session/session-checkpoint-policy), `timeout-policy` | | `tools/post-execute` | `waterfall` | [`packages/core/tools/src/index.ts:175`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`waterfall`) | [`hooks-claude-code`](../packages/hooks/hooks-claude-code), [`hooks-codex`](../packages/hooks/hooks-codex), [`repeat-tool-reminder`](../packages/guard/repeat-tool-reminder), [`spill-policy`](../packages/spill/spill-policy), [`tool-fs-search`](../packages/fs/tool-fs-search) | diff --git a/docs/module-graph.i18n.yaml b/docs/module-graph.i18n.yaml index a7edea20b2..a467f4aea6 100644 --- a/docs/module-graph.i18n.yaml +++ b/docs/module-graph.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/module-graph.md -module-graph.md: a2497ed002c6eec55ec35e1d7e9953feb2ed0ee5 -module-graph.zh.md: 230d67f4c07f9715d4b5157c1539edd09f9221c3 +module-graph.md: 9d192404f4949bcf6e5c4222f0b898cc52362bd8 +module-graph.zh.md: bbdfd5642238644939fd08c02b9786310bad63e2 diff --git a/docs/module-graph.md b/docs/module-graph.md index a2497ed002..9d192404f4 100644 --- a/docs/module-graph.md +++ b/docs/module-graph.md @@ -570,14 +570,6 @@ flowchart TD pkg_jobs --> pkg_brand pkg_jobs --> pkg_invariants pkg_jobs --> pkg_session - pkg_agent_presets --> pkg_agent - pkg_agent_presets --> pkg_atomic_write - pkg_agent_presets --> pkg_home_paths - pkg_agent_presets --> pkg_invariants - pkg_agent_presets --> pkg_scope - pkg_agent_presets --> pkg_session - pkg_agent_presets --> pkg_settings - pkg_agent_presets --> pkg_system_prompt pkg_sandbox_local --> pkg_invariants pkg_sandbox_local --> pkg_llm pkg_sandbox_local --> pkg_sandbox @@ -655,11 +647,6 @@ flowchart TD pkg_llm_pi_ai --> pkg_llm pkg_llm_pi_ai --> pkg_settings pkg_llm_pi_ai --> pkg_timeout - pkg_plugin_package_inventory_deepseek --> pkg_agent - pkg_plugin_package_inventory_deepseek --> pkg_agent_presets - pkg_plugin_package_inventory_deepseek --> pkg_deepseek_llm_api_extensions - pkg_plugin_package_inventory_deepseek --> pkg_invariants - pkg_plugin_package_inventory_deepseek --> pkg_session pkg_tools --> pkg_agent pkg_tools --> pkg_code_runtime pkg_tools --> pkg_invariants @@ -710,8 +697,6 @@ flowchart TD pkg_command_feedback --> pkg_invariants pkg_command_feedback --> pkg_session pkg_command_feedback --> pkg_session_telemetry - pkg_host_apiproxy --> pkg_agent_presets - pkg_host_apiproxy --> pkg_invariants pkg_permission_presets --> pkg_commands pkg_permission_presets --> pkg_invariants pkg_permission_presets --> pkg_sandbox @@ -812,21 +797,6 @@ flowchart TD pkg_tool_skill --> pkg_llm pkg_tool_skill --> pkg_skill pkg_tool_skill --> pkg_tools - pkg_subagent --> pkg_agent - pkg_subagent --> pkg_agent_presets - pkg_subagent --> pkg_brand - pkg_subagent --> pkg_invariants - pkg_subagent --> pkg_jobs - pkg_subagent --> pkg_llm - pkg_subagent --> pkg_sandbox - pkg_subagent --> pkg_sandbox_policy - pkg_subagent --> pkg_scope - pkg_subagent --> pkg_session - pkg_subagent --> pkg_session_persistence - pkg_subagent --> pkg_session_projection - pkg_subagent --> pkg_session_projection_cache - pkg_subagent --> pkg_tools - pkg_subagent --> pkg_user_approval pkg_tool_web --> pkg_invariants pkg_tool_web --> pkg_llm pkg_tool_web --> pkg_system_prompt @@ -874,10 +844,6 @@ flowchart TD pkg_file_reference_local --> pkg_invariants pkg_file_reference_local --> pkg_system_prompt pkg_file_reference_local --> pkg_tools - pkg_experimental_webworker_runtime --> pkg_client_modules - pkg_experimental_webworker_runtime --> pkg_host_apiproxy - pkg_experimental_webworker_runtime --> pkg_host_webserver - pkg_experimental_webworker_runtime --> pkg_invariants pkg_cordis_host_runner --> pkg_agent pkg_cordis_host_runner --> pkg_brand pkg_cordis_host_runner --> pkg_invariants @@ -917,6 +883,15 @@ flowchart TD pkg_mcp_client --> pkg_subprocess pkg_mcp_client --> pkg_timeout pkg_mcp_client --> pkg_tools + pkg_agent_presets --> pkg_agent + pkg_agent_presets --> pkg_atomic_write + pkg_agent_presets --> pkg_home_paths + pkg_agent_presets --> pkg_invariants + pkg_agent_presets --> pkg_scope + pkg_agent_presets --> pkg_session + pkg_agent_presets --> pkg_settings + pkg_agent_presets --> pkg_system_prompt + pkg_agent_presets --> pkg_tools pkg_schedule --> pkg_agent pkg_schedule --> pkg_brand pkg_schedule --> pkg_invariants @@ -990,6 +965,89 @@ flowchart TD pkg_llm_replay --> pkg_invariants pkg_llm_replay --> pkg_llm pkg_llm_replay --> pkg_session + pkg_tool_workflow --> pkg_agent + pkg_tool_workflow --> pkg_invariants + pkg_tool_workflow --> pkg_llm + pkg_tool_workflow --> pkg_session + pkg_tool_workflow --> pkg_system_prompt + pkg_tool_workflow --> pkg_tools + pkg_tool_workflow --> pkg_workflow + pkg_plugin_package_inventory_deepseek --> pkg_agent + pkg_plugin_package_inventory_deepseek --> pkg_agent_presets + pkg_plugin_package_inventory_deepseek --> pkg_deepseek_llm_api_extensions + pkg_plugin_package_inventory_deepseek --> pkg_invariants + pkg_plugin_package_inventory_deepseek --> pkg_session + pkg_subagent --> pkg_agent + pkg_subagent --> pkg_agent_presets + pkg_subagent --> pkg_brand + pkg_subagent --> pkg_invariants + pkg_subagent --> pkg_jobs + pkg_subagent --> pkg_llm + pkg_subagent --> pkg_sandbox + pkg_subagent --> pkg_sandbox_policy + pkg_subagent --> pkg_scope + pkg_subagent --> pkg_session + pkg_subagent --> pkg_session_persistence + pkg_subagent --> pkg_session_projection + pkg_subagent --> pkg_session_projection_cache + pkg_subagent --> pkg_tools + pkg_subagent --> pkg_user_approval + pkg_session_query --> pkg_brand + pkg_session_query --> pkg_invariants + pkg_session_query --> pkg_llm + pkg_session_query --> pkg_session + pkg_session_query --> pkg_session_persistence + pkg_session_query --> pkg_session_title + pkg_session_query --> pkg_tool_todo + pkg_acp --> pkg_agent + pkg_acp --> pkg_attachment + pkg_acp --> pkg_invariants + pkg_acp --> pkg_llm + pkg_acp --> pkg_mcp_client + pkg_acp --> pkg_session + pkg_acp --> pkg_session_persistence + pkg_acp --> pkg_token_meter + pkg_acp --> pkg_user_approval + pkg_web_app --> pkg_invariants + pkg_web_app --> pkg_shell_env + pkg_web_app --> pkg_system_prompt + pkg_compaction_tool_result_pruner --> pkg_compaction + pkg_compaction_tool_result_pruner --> pkg_invariants + pkg_compaction_tool_result_pruner --> pkg_llm + pkg_compaction_tool_result_pruner --> pkg_session + pkg_compaction_tool_result_pruner --> pkg_token_meter + pkg_tool_cordis --> pkg_agent + pkg_tool_cordis --> pkg_cordis_host_runner + pkg_tool_cordis --> pkg_invariants + pkg_tool_cordis --> pkg_llm + pkg_tool_cordis --> pkg_scope + pkg_tool_cordis --> pkg_session + pkg_tool_cordis --> pkg_system_prompt + pkg_tool_cordis --> pkg_tools + pkg_host_apiproxy --> pkg_agent_presets + pkg_host_apiproxy --> pkg_invariants + pkg_tool_bash --> pkg_agent + pkg_tool_bash --> pkg_invariants + pkg_tool_bash --> pkg_jobs + pkg_tool_bash --> pkg_llm + pkg_tool_bash --> pkg_sandbox + pkg_tool_bash --> pkg_sandbox_policy + pkg_tool_bash --> pkg_shell + pkg_tool_bash --> pkg_shell_env + pkg_tool_bash --> pkg_system_prompt + pkg_tool_bash --> pkg_tools + pkg_tool_bash --> pkg_user_approval + pkg_tool_pwsh --> pkg_agent + pkg_tool_pwsh --> pkg_invariants + pkg_tool_pwsh --> pkg_jobs + pkg_tool_pwsh --> pkg_llm + pkg_tool_pwsh --> pkg_sandbox + pkg_tool_pwsh --> pkg_sandbox_policy + pkg_tool_pwsh --> pkg_shell + pkg_tool_pwsh --> pkg_shell_env + pkg_tool_pwsh --> pkg_system_prompt + pkg_tool_pwsh --> pkg_tools + pkg_tool_pwsh --> pkg_user_approval pkg_webhook --> pkg_agent pkg_webhook --> pkg_agent_default_model pkg_webhook --> pkg_agent_presets @@ -1000,13 +1058,6 @@ flowchart TD pkg_webhook --> pkg_session pkg_webhook --> pkg_session_title pkg_webhook --> pkg_workspace - pkg_tool_workflow --> pkg_agent - pkg_tool_workflow --> pkg_invariants - pkg_tool_workflow --> pkg_llm - pkg_tool_workflow --> pkg_session - pkg_tool_workflow --> pkg_system_prompt - pkg_tool_workflow --> pkg_tools - pkg_tool_workflow --> pkg_workflow pkg_subagent_acp --> pkg_agent pkg_subagent_acp --> pkg_invariants pkg_subagent_acp --> pkg_llm @@ -1037,6 +1088,9 @@ flowchart TD pkg_tool_subagent --> pkg_invariants pkg_tool_subagent --> pkg_jobs pkg_tool_subagent --> pkg_llm + pkg_tool_subagent --> pkg_scope + pkg_tool_subagent --> pkg_session + pkg_tool_subagent --> pkg_settings pkg_tool_subagent --> pkg_subagent pkg_tool_subagent --> pkg_system_prompt pkg_tool_subagent --> pkg_tools @@ -1058,107 +1112,6 @@ flowchart TD pkg_hooks_claude_code --> pkg_session_persistence pkg_hooks_claude_code --> pkg_subagent pkg_hooks_claude_code --> pkg_tools - pkg_session_query --> pkg_brand - pkg_session_query --> pkg_invariants - pkg_session_query --> pkg_llm - pkg_session_query --> pkg_session - pkg_session_query --> pkg_session_persistence - pkg_session_query --> pkg_session_title - pkg_session_query --> pkg_tool_todo - pkg_acp --> pkg_agent - pkg_acp --> pkg_attachment - pkg_acp --> pkg_invariants - pkg_acp --> pkg_llm - pkg_acp --> pkg_mcp_client - pkg_acp --> pkg_session - pkg_acp --> pkg_session_persistence - pkg_acp --> pkg_token_meter - pkg_acp --> pkg_user_approval - pkg_web_app --> pkg_invariants - pkg_web_app --> pkg_shell_env - pkg_web_app --> pkg_system_prompt - pkg_client_connection --> pkg_attachment - pkg_client_connection --> pkg_commands - pkg_client_connection --> pkg_host_apiproxy - pkg_client_connection --> pkg_host_webserver - pkg_client_connection --> pkg_invariants - pkg_client_connection --> pkg_llm - pkg_client_connection --> pkg_session - pkg_client_connection --> pkg_tool_todo - pkg_compaction_tool_result_pruner --> pkg_compaction - pkg_compaction_tool_result_pruner --> pkg_invariants - pkg_compaction_tool_result_pruner --> pkg_llm - pkg_compaction_tool_result_pruner --> pkg_session - pkg_compaction_tool_result_pruner --> pkg_token_meter - pkg_experimental_agent_team --> pkg_agent - pkg_experimental_agent_team --> pkg_brand - pkg_experimental_agent_team --> pkg_invariants - pkg_experimental_agent_team --> pkg_llm - pkg_experimental_agent_team --> pkg_session - pkg_experimental_agent_team --> pkg_session_persistence - pkg_experimental_agent_team --> pkg_subagent - pkg_tool_cordis --> pkg_agent - pkg_tool_cordis --> pkg_cordis_host_runner - pkg_tool_cordis --> pkg_invariants - pkg_tool_cordis --> pkg_llm - pkg_tool_cordis --> pkg_scope - pkg_tool_cordis --> pkg_session - pkg_tool_cordis --> pkg_system_prompt - pkg_tool_cordis --> pkg_tools - pkg_sdk_protocol --> pkg_invariants - pkg_sdk_protocol --> pkg_llm - pkg_sdk_protocol --> pkg_session - pkg_sdk_protocol --> pkg_subagent - pkg_tool_bash --> pkg_agent - pkg_tool_bash --> pkg_invariants - pkg_tool_bash --> pkg_jobs - pkg_tool_bash --> pkg_llm - pkg_tool_bash --> pkg_sandbox - pkg_tool_bash --> pkg_sandbox_policy - pkg_tool_bash --> pkg_shell - pkg_tool_bash --> pkg_shell_env - pkg_tool_bash --> pkg_system_prompt - pkg_tool_bash --> pkg_tools - pkg_tool_bash --> pkg_user_approval - pkg_tool_pwsh --> pkg_agent - pkg_tool_pwsh --> pkg_invariants - pkg_tool_pwsh --> pkg_jobs - pkg_tool_pwsh --> pkg_llm - pkg_tool_pwsh --> pkg_sandbox - pkg_tool_pwsh --> pkg_sandbox_policy - pkg_tool_pwsh --> pkg_shell - pkg_tool_pwsh --> pkg_shell_env - pkg_tool_pwsh --> pkg_system_prompt - pkg_tool_pwsh --> pkg_tools - pkg_tool_pwsh --> pkg_user_approval - pkg_webhook_github --> pkg_credentials - pkg_webhook_github --> pkg_host_webserver - pkg_webhook_github --> pkg_invariants - pkg_webhook_github --> pkg_session - pkg_webhook_github --> pkg_webhook - pkg_tool_ralph --> pkg_agent - pkg_tool_ralph --> pkg_invariants - pkg_tool_ralph --> pkg_llm - pkg_tool_ralph --> pkg_subagent - pkg_tool_ralph --> pkg_system_prompt - pkg_tool_ralph --> pkg_tools - pkg_tool_ralph --> pkg_workflow - pkg_workflow_worker_thread --> pkg_agent - pkg_workflow_worker_thread --> pkg_brand - pkg_workflow_worker_thread --> pkg_invariants - pkg_workflow_worker_thread --> pkg_llm - pkg_workflow_worker_thread --> pkg_session - pkg_workflow_worker_thread --> pkg_subagent - pkg_workflow_worker_thread --> pkg_tools - pkg_workflow_worker_thread --> pkg_workflow - pkg_subagent_fork_in_process --> pkg_agent - pkg_subagent_fork_in_process --> pkg_invariants - pkg_subagent_fork_in_process --> pkg_session - pkg_subagent_fork_in_process --> pkg_subagent - pkg_subagent_fork_in_process --> pkg_subagent_in_process_driver - pkg_subagent_spawn_in_process --> pkg_invariants - pkg_subagent_spawn_in_process --> pkg_subagent - pkg_subagent_spawn_in_process --> pkg_subagent_in_process_driver pkg_session_query_sqlite --> pkg_invariants pkg_session_query_sqlite --> pkg_session pkg_session_query_sqlite --> pkg_session_persistence @@ -1170,11 +1123,14 @@ flowchart TD pkg_tool_session_query --> pkg_system_prompt pkg_tool_session_query --> pkg_timeout pkg_tool_session_query --> pkg_tools - pkg_api_gateway --> pkg_brand - pkg_api_gateway --> pkg_client_connection - pkg_api_gateway --> pkg_host_webserver - pkg_api_gateway --> pkg_invariants - pkg_api_gateway --> pkg_typert_registry + pkg_client_connection --> pkg_attachment + pkg_client_connection --> pkg_commands + pkg_client_connection --> pkg_host_apiproxy + pkg_client_connection --> pkg_host_webserver + pkg_client_connection --> pkg_invariants + pkg_client_connection --> pkg_llm + pkg_client_connection --> pkg_session + pkg_client_connection --> pkg_tool_todo pkg_compaction_basic --> pkg_agent pkg_compaction_basic --> pkg_commands pkg_compaction_basic --> pkg_compaction @@ -1213,6 +1169,54 @@ flowchart TD pkg_agent_spine_demo --> pkg_tool_jobs pkg_agent_spine_demo --> pkg_tool_skill pkg_agent_spine_demo --> pkg_tools + pkg_experimental_agent_team --> pkg_agent + pkg_experimental_agent_team --> pkg_brand + pkg_experimental_agent_team --> pkg_invariants + pkg_experimental_agent_team --> pkg_llm + pkg_experimental_agent_team --> pkg_session + pkg_experimental_agent_team --> pkg_session_persistence + pkg_experimental_agent_team --> pkg_subagent + pkg_experimental_webworker_runtime --> pkg_client_modules + pkg_experimental_webworker_runtime --> pkg_host_apiproxy + pkg_experimental_webworker_runtime --> pkg_host_webserver + pkg_experimental_webworker_runtime --> pkg_invariants + pkg_sdk_protocol --> pkg_invariants + pkg_sdk_protocol --> pkg_llm + pkg_sdk_protocol --> pkg_session + pkg_sdk_protocol --> pkg_subagent + pkg_webhook_github --> pkg_credentials + pkg_webhook_github --> pkg_host_webserver + pkg_webhook_github --> pkg_invariants + pkg_webhook_github --> pkg_session + pkg_webhook_github --> pkg_webhook + pkg_tool_ralph --> pkg_agent + pkg_tool_ralph --> pkg_invariants + pkg_tool_ralph --> pkg_llm + pkg_tool_ralph --> pkg_subagent + pkg_tool_ralph --> pkg_system_prompt + pkg_tool_ralph --> pkg_tools + pkg_tool_ralph --> pkg_workflow + pkg_workflow_worker_thread --> pkg_agent + pkg_workflow_worker_thread --> pkg_brand + pkg_workflow_worker_thread --> pkg_invariants + pkg_workflow_worker_thread --> pkg_llm + pkg_workflow_worker_thread --> pkg_session + pkg_workflow_worker_thread --> pkg_subagent + pkg_workflow_worker_thread --> pkg_tools + pkg_workflow_worker_thread --> pkg_workflow + pkg_subagent_fork_in_process --> pkg_agent + pkg_subagent_fork_in_process --> pkg_invariants + pkg_subagent_fork_in_process --> pkg_session + pkg_subagent_fork_in_process --> pkg_subagent + pkg_subagent_fork_in_process --> pkg_subagent_in_process_driver + pkg_subagent_spawn_in_process --> pkg_invariants + pkg_subagent_spawn_in_process --> pkg_subagent + pkg_subagent_spawn_in_process --> pkg_subagent_in_process_driver + pkg_api_gateway --> pkg_brand + pkg_api_gateway --> pkg_client_connection + pkg_api_gateway --> pkg_host_webserver + pkg_api_gateway --> pkg_invariants + pkg_api_gateway --> pkg_typert_registry pkg_experimental_tool_agent_team --> pkg_agent pkg_experimental_tool_agent_team --> pkg_experimental_agent_team pkg_experimental_tool_agent_team --> pkg_invariants @@ -1731,7 +1735,6 @@ flowchart TD | [`user-approval`](../packages/interaction/user-approval) | `interaction` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt) | | [`user-questions`](../packages/interaction/user-questions) | `interaction` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope) | | [`jobs`](../packages/jobs/jobs) | `jobs` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session) | -| [`agent-presets`](../packages/preset/agent-presets) | `preset` | [`agent`](../packages/core/agent), [`atomic-write`](../packages/util/atomic-write), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`settings`](../packages/settings/settings), [`system-prompt`](../packages/core/system-prompt) | | [`sandbox-local`](../packages/sandbox/sandbox-local) | `sandbox` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`session`](../packages/core/session) | | [`sandbox-policy`](../packages/sandbox/sandbox-policy) | `sandbox` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`sandbox`](../packages/sandbox/sandbox), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt) | | [`session-persistence-jsonl`](../packages/session/session-persistence-jsonl) | `session` | [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence) | @@ -1747,7 +1750,6 @@ flowchart TD | [`workspace`](../packages/workspace/workspace) | `workspace` | [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`storage`](../packages/storage/storage), [`storage-domain`](../packages/storage/storage-domain) | | [`llm-deepseek`](../packages/llm/llm-deepseek) | `llm` | [`anonymous-user-id`](../packages/identity/anonymous-user-id), [`atomic-write`](../packages/util/atomic-write), [`attachment`](../packages/attachment/attachment), [`brand`](../packages/util/brand), [`credentials`](../packages/credentials/credentials), [`deepseek-llm-api-extensions`](../packages/llm/deepseek-llm-api-extensions), [`fs`](../packages/fs/fs), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`launch-environment`](../packages/util/launch-environment), [`llm`](../packages/llm/llm), [`settings`](../packages/settings/settings), [`timeout`](../packages/util/timeout) | | [`llm-pi-ai`](../packages/llm/llm-pi-ai) | `llm` | [`attachment`](../packages/attachment/attachment), [`authorization`](../packages/credentials/authorization), [`credentials`](../packages/credentials/credentials), [`fs`](../packages/fs/fs), [`invariants`](../packages/runtime-diagnostics/invariants), [`launch-environment`](../packages/util/launch-environment), [`llm`](../packages/llm/llm), [`settings`](../packages/settings/settings), [`timeout`](../packages/util/timeout) | -| [`plugin-package-inventory-deepseek`](../packages/llm/plugin-package-inventory-deepseek) | `llm` | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`deepseek-llm-api-extensions`](../packages/llm/deepseek-llm-api-extensions), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session) | | [`tools`](../packages/core/tools) | `core` | [`agent`](../packages/core/agent), [`code-runtime`](../packages/code-runtime/code-runtime), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`user-approval`](../packages/interaction/user-approval) | | [`command-goal`](../packages/goal/command-goal) | `goal` | [`commands`](../packages/interaction/commands), [`goal`](../packages/goal/goal), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm) | | [`goal-round-driver`](../packages/goal/goal-round-driver) | `goal` | [`agent`](../packages/core/agent), [`goal`](../packages/goal/goal), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session) | @@ -1760,7 +1762,6 @@ flowchart TD | [`tmux-context`](../packages/context/tmux-context) | `context` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`shell`](../packages/shell/shell) | | [`fs-e2b`](../packages/e2b/fs-e2b) | `e2b` | [`e2b`](../packages/e2b/e2b), [`fs`](../packages/fs/fs), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`command-feedback`](../packages/feedback/command-feedback) | `feedback` | [`anonymous-user-id`](../packages/identity/anonymous-user-id), [`commands`](../packages/interaction/commands), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-telemetry`](../packages/session/session-telemetry) | -| [`host-apiproxy`](../packages/host/apiproxy) | `host` | [`agent-presets`](../packages/preset/agent-presets), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`permission-presets`](../packages/interaction/permission-presets) | `interaction` | [`commands`](../packages/interaction/commands), [`invariants`](../packages/runtime-diagnostics/invariants), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`session`](../packages/core/session), [`session-projection`](../packages/session/session-projection), [`settings`](../packages/settings/settings), [`shell`](../packages/shell/shell), [`user-approval`](../packages/interaction/user-approval) | | [`jobs-local`](../packages/jobs/jobs-local) | `jobs` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`scope`](../packages/core/scope), [`timeout`](../packages/util/timeout) | | [`lsp-stdio`](../packages/lsp/lsp-stdio) | `lsp` | [`brand`](../packages/util/brand), [`fs`](../packages/fs/fs), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`lsp`](../packages/lsp/lsp), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout) | @@ -1777,7 +1778,6 @@ flowchart TD | [`tool-fs-search`](../packages/fs/tool-fs-search) | `fs` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`session`](../packages/core/session), [`spill`](../packages/spill/spill), [`subprocess`](../packages/subprocess/subprocess), [`system-prompt`](../packages/core/system-prompt), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | | [`tool-str-replace-editor`](../packages/fs/tool-str-replace-editor) | `fs` | [`fs`](../packages/fs/fs), [`invariants`](../packages/runtime-diagnostics/invariants), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`tools`](../packages/core/tools) | | [`tool-skill`](../packages/skill/tool-skill) | `skill` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`skill`](../packages/skill/skill), [`tools`](../packages/core/tools) | -| [`subagent`](../packages/subagent/subagent) | `subagent` | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-projection`](../packages/session/session-projection), [`session-projection-cache`](../packages/session/session-projection-cache), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | | [`tool-web`](../packages/web/tool-web) | `web` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`web`](../packages/web/web) | | [`spill-policy`](../packages/spill/spill-policy) | `spill` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`session`](../packages/core/session), [`spill`](../packages/spill/spill), [`tools`](../packages/core/tools) | | [`tool-todo`](../packages/todo/tool-todo) | `todo` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-projection`](../packages/session/session-projection), [`tools`](../packages/core/tools) | @@ -1786,7 +1786,6 @@ flowchart TD | [`command-compact`](../packages/compaction/command-compact) | `compaction` | [`commands`](../packages/interaction/commands), [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`agent-instructions`](../packages/context/agent-instructions) | `context` | [`agent`](../packages/core/agent), [`fs`](../packages/fs/fs), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`tools`](../packages/core/tools) | | [`file-reference-local`](../packages/context/file-reference-local) | `context` | [`agent`](../packages/core/agent), [`file-reference`](../packages/context/file-reference), [`invariants`](../packages/runtime-diagnostics/invariants), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | -| [`experimental-webworker-runtime`](../packages/experimental/webworker-runtime) | `experimental` | [`client-modules`](../packages/client/modules), [`host-apiproxy`](../packages/host/apiproxy), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`cordis-host-runner`](../packages/extensions/cordis-host-runner) | `extensions` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`tools`](../packages/core/tools), [`typert-protocol`](../packages/typert/protocol) | | [`repeat-tool-reminder`](../packages/guard/repeat-tool-reminder) | `guard` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`tools`](../packages/core/tools) | | [`tool-call-timeout-policy`](../packages/guard/timeout-policy) | `guard` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | @@ -1794,6 +1793,7 @@ flowchart TD | [`tool-jobs`](../packages/jobs/tool-jobs) | `jobs` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`tool-lsp`](../packages/lsp/tool-lsp) | `lsp` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`lsp`](../packages/lsp/lsp), [`system-prompt`](../packages/core/system-prompt), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | | [`mcp-client`](../packages/mcp/mcp-client) | `mcp` | [`attachment`](../packages/attachment/attachment), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | +| [`agent-presets`](../packages/preset/agent-presets) | `preset` | [`agent`](../packages/core/agent), [`atomic-write`](../packages/util/atomic-write), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`settings`](../packages/settings/settings), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`schedule`](../packages/schedule/schedule) | `schedule` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`tools`](../packages/core/tools) | | [`session-checkpoint-policy`](../packages/session/session-checkpoint-policy) | `session` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`tools`](../packages/core/tools) | | [`session-telemetry-otel`](../packages/session/session-telemetry-otel) | `session` | [`anonymous-user-id`](../packages/identity/anonymous-user-id), [`command-feedback`](../packages/feedback/command-feedback), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-telemetry`](../packages/session/session-telemetry) | @@ -1807,37 +1807,41 @@ flowchart TD | [`tool-terminal`](../packages/terminal/tool-terminal) | `terminal` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`system-prompt`](../packages/core/system-prompt), [`terminal`](../packages/terminal/terminal), [`tools`](../packages/core/tools) | | [`agent-loop-testkit`](../packages/test-support/agent-loop-testkit) | `test-support` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`llm-replay`](../packages/test-support/llm-replay) | `test-support` | [`compaction`](../packages/compaction/compaction), [`deepseek-llm-api-extensions`](../packages/llm/deepseek-llm-api-extensions), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session) | -| [`webhook`](../packages/webhook/webhook) | `webhook` | [`agent`](../packages/core/agent), [`agent-default-model`](../packages/core/agent-default-model), [`agent-presets`](../packages/preset/agent-presets), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`permission-presets`](../packages/interaction/permission-presets), [`session`](../packages/core/session), [`session-title`](../packages/session/session-title), [`workspace`](../packages/workspace/workspace) | | [`tool-workflow`](../packages/workflow/tool-workflow) | `workflow` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`workflow`](../packages/workflow/workflow) | +| [`plugin-package-inventory-deepseek`](../packages/llm/plugin-package-inventory-deepseek) | `llm` | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`deepseek-llm-api-extensions`](../packages/llm/deepseek-llm-api-extensions), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session) | +| [`subagent`](../packages/subagent/subagent) | `subagent` | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-projection`](../packages/session/session-projection), [`session-projection-cache`](../packages/session/session-projection-cache), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | +| [`session-query`](../packages/session-query/session-query) | `session-query` | [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-title`](../packages/session/session-title), [`tool-todo`](../packages/todo/tool-todo) | +| [`acp`](../packages/acp/acp) | `acp` | [`agent`](../packages/core/agent), [`attachment`](../packages/attachment/attachment), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`mcp-client`](../packages/mcp/mcp-client), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`token-meter`](../packages/llm/token-meter), [`user-approval`](../packages/interaction/user-approval) | +| [`web-app`](../packages/bundle/web-app) | `bundle` | [`invariants`](../packages/runtime-diagnostics/invariants), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt) | +| [`compaction-tool-result-pruner`](../packages/compaction/compaction-tool-result-pruner) | `compaction` | [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`token-meter`](../packages/llm/token-meter) | +| [`tool-cordis`](../packages/extensions/tool-cordis) | `extensions` | [`agent`](../packages/core/agent), [`cordis-host-runner`](../packages/extensions/cordis-host-runner), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | +| [`host-apiproxy`](../packages/host/apiproxy) | `host` | [`agent-presets`](../packages/preset/agent-presets), [`invariants`](../packages/runtime-diagnostics/invariants) | +| [`tool-bash`](../packages/shell/tool-bash) | `shell` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`shell`](../packages/shell/shell), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | +| [`tool-pwsh`](../packages/shell/tool-pwsh) | `shell` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`shell`](../packages/shell/shell), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | +| [`webhook`](../packages/webhook/webhook) | `webhook` | [`agent`](../packages/core/agent), [`agent-default-model`](../packages/core/agent-default-model), [`agent-presets`](../packages/preset/agent-presets), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`permission-presets`](../packages/interaction/permission-presets), [`session`](../packages/core/session), [`session-title`](../packages/session/session-title), [`workspace`](../packages/workspace/workspace) | | [`subagent-acp`](../packages/subagent/subagent-acp) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout) | | [`subagent-claude-code`](../packages/subagent/subagent-claude-code) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout) | | [`subagent-codex`](../packages/subagent/subagent-codex) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout) | | [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | -| [`tool-subagent`](../packages/subagent/tool-subagent) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | +| [`tool-subagent`](../packages/subagent/tool-subagent) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`settings`](../packages/settings/settings), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`tool-subagent-control`](../packages/subagent/tool-subagent-control) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`tools`](../packages/core/tools) | | [`tool-subagent-report`](../packages/subagent/tool-subagent-report) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`hooks-claude-code`](../packages/hooks/hooks-claude-code) | `hooks` | [`agent`](../packages/core/agent), [`hook-protocol`](../packages/hooks/hook-protocol), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`subagent`](../packages/subagent/subagent), [`tools`](../packages/core/tools) | -| [`session-query`](../packages/session-query/session-query) | `session-query` | [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-title`](../packages/session/session-title), [`tool-todo`](../packages/todo/tool-todo) | -| [`acp`](../packages/acp/acp) | `acp` | [`agent`](../packages/core/agent), [`attachment`](../packages/attachment/attachment), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`mcp-client`](../packages/mcp/mcp-client), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`token-meter`](../packages/llm/token-meter), [`user-approval`](../packages/interaction/user-approval) | -| [`web-app`](../packages/bundle/web-app) | `bundle` | [`invariants`](../packages/runtime-diagnostics/invariants), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt) | +| [`session-query-sqlite`](../packages/session-query/session-query-sqlite) | `session-query` | [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-query`](../packages/session-query/session-query) | +| [`tool-session-query`](../packages/session-query/tool-session-query) | `session-query` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-query`](../packages/session-query/session-query), [`system-prompt`](../packages/core/system-prompt), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | | [`client-connection`](../packages/client/connection) | `client` | [`attachment`](../packages/attachment/attachment), [`commands`](../packages/interaction/commands), [`host-apiproxy`](../packages/host/apiproxy), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`tool-todo`](../packages/todo/tool-todo) | -| [`compaction-tool-result-pruner`](../packages/compaction/compaction-tool-result-pruner) | `compaction` | [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`token-meter`](../packages/llm/token-meter) | +| [`compaction-basic`](../packages/compaction/compaction-basic) | `compaction` | [`agent`](../packages/core/agent), [`commands`](../packages/interaction/commands), [`compaction`](../packages/compaction/compaction), [`compaction-tool-result-pruner`](../packages/compaction/compaction-tool-result-pruner), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`token-meter`](../packages/llm/token-meter) | +| [`session-reference`](../packages/context/session-reference) | `context` | [`agent`](../packages/core/agent), [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`session`](../packages/core/session), [`session-query`](../packages/session-query/session-query), [`typert-protocol`](../packages/typert/protocol) | +| [`agent-spine-demo`](../packages/examples/agent-spine-demo) | `examples` | [`agent`](../packages/core/agent), [`agent-instructions`](../packages/context/agent-instructions), [`agent-loop`](../packages/core/agent-loop), [`goal`](../packages/goal/goal), [`goal-round-driver`](../packages/goal/goal-round-driver), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs-local`](../packages/jobs/jobs-local), [`llm`](../packages/llm/llm), [`llm-retry`](../packages/llm/llm-retry), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`session-title`](../packages/session/session-title), [`shell-env`](../packages/shell/shell-env), [`skill`](../packages/skill/skill), [`skill-filesystem`](../packages/skill/skill-filesystem), [`system-prompt`](../packages/core/system-prompt), [`tool-bash`](../packages/shell/tool-bash), [`tool-goal`](../packages/goal/tool-goal), [`tool-jobs`](../packages/jobs/tool-jobs), [`tool-skill`](../packages/skill/tool-skill), [`tools`](../packages/core/tools) | | [`experimental-agent-team`](../packages/experimental/agent-team) | `experimental` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`subagent`](../packages/subagent/subagent) | -| [`tool-cordis`](../packages/extensions/tool-cordis) | `extensions` | [`agent`](../packages/core/agent), [`cordis-host-runner`](../packages/extensions/cordis-host-runner), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | +| [`experimental-webworker-runtime`](../packages/experimental/webworker-runtime) | `experimental` | [`client-modules`](../packages/client/modules), [`host-apiproxy`](../packages/host/apiproxy), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`sdk-protocol`](../packages/sdk/protocol) | `sdk` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent) | -| [`tool-bash`](../packages/shell/tool-bash) | `shell` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`shell`](../packages/shell/shell), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | -| [`tool-pwsh`](../packages/shell/tool-pwsh) | `shell` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`shell`](../packages/shell/shell), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | | [`webhook-github`](../packages/webhook/webhook-github) | `webhook` | [`credentials`](../packages/credentials/credentials), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`webhook`](../packages/webhook/webhook) | | [`tool-ralph`](../packages/workflow/tool-ralph) | `workflow` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`workflow`](../packages/workflow/workflow) | | [`workflow-worker-thread`](../packages/workflow/workflow-worker-thread) | `workflow` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`tools`](../packages/core/tools), [`workflow`](../packages/workflow/workflow) | | [`subagent-fork-in-process`](../packages/subagent/subagent-fork-in-process) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver) | | [`subagent-spawn-in-process`](../packages/subagent/subagent-spawn-in-process) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`subagent`](../packages/subagent/subagent), [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver) | -| [`session-query-sqlite`](../packages/session-query/session-query-sqlite) | `session-query` | [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-query`](../packages/session-query/session-query) | -| [`tool-session-query`](../packages/session-query/tool-session-query) | `session-query` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-query`](../packages/session-query/session-query), [`system-prompt`](../packages/core/system-prompt), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | | [`api-gateway`](../packages/api/gateway) | `api` | [`brand`](../packages/util/brand), [`client-connection`](../packages/client/connection), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants), [`typert-registry`](../packages/typert/registry) | -| [`compaction-basic`](../packages/compaction/compaction-basic) | `compaction` | [`agent`](../packages/core/agent), [`commands`](../packages/interaction/commands), [`compaction`](../packages/compaction/compaction), [`compaction-tool-result-pruner`](../packages/compaction/compaction-tool-result-pruner), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`token-meter`](../packages/llm/token-meter) | -| [`session-reference`](../packages/context/session-reference) | `context` | [`agent`](../packages/core/agent), [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`session`](../packages/core/session), [`session-query`](../packages/session-query/session-query), [`typert-protocol`](../packages/typert/protocol) | -| [`agent-spine-demo`](../packages/examples/agent-spine-demo) | `examples` | [`agent`](../packages/core/agent), [`agent-instructions`](../packages/context/agent-instructions), [`agent-loop`](../packages/core/agent-loop), [`goal`](../packages/goal/goal), [`goal-round-driver`](../packages/goal/goal-round-driver), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs-local`](../packages/jobs/jobs-local), [`llm`](../packages/llm/llm), [`llm-retry`](../packages/llm/llm-retry), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`session-title`](../packages/session/session-title), [`shell-env`](../packages/shell/shell-env), [`skill`](../packages/skill/skill), [`skill-filesystem`](../packages/skill/skill-filesystem), [`system-prompt`](../packages/core/system-prompt), [`tool-bash`](../packages/shell/tool-bash), [`tool-goal`](../packages/goal/tool-goal), [`tool-jobs`](../packages/jobs/tool-jobs), [`tool-skill`](../packages/skill/tool-skill), [`tools`](../packages/core/tools) | | [`experimental-tool-agent-team`](../packages/experimental/tool-agent-team) | `experimental` | [`agent`](../packages/core/agent), [`experimental-agent-team`](../packages/experimental/agent-team), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`sdk-client`](../packages/sdk/client) | `sdk` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`sdk-protocol`](../packages/sdk/protocol), [`session`](../packages/core/session) | | [`sdk-jsonrpc-server`](../packages/sdk/server) | `sdk` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`llm-deepseek`](../packages/llm/llm-deepseek), [`scope`](../packages/core/scope), [`sdk-protocol`](../packages/sdk/protocol), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent) | diff --git a/docs/module-graph.zh.md b/docs/module-graph.zh.md index 230d67f4c0..bbdfd56422 100644 --- a/docs/module-graph.zh.md +++ b/docs/module-graph.zh.md @@ -572,14 +572,6 @@ flowchart TD pkg_jobs --> pkg_brand pkg_jobs --> pkg_invariants pkg_jobs --> pkg_session - pkg_agent_presets --> pkg_agent - pkg_agent_presets --> pkg_atomic_write - pkg_agent_presets --> pkg_home_paths - pkg_agent_presets --> pkg_invariants - pkg_agent_presets --> pkg_scope - pkg_agent_presets --> pkg_session - pkg_agent_presets --> pkg_settings - pkg_agent_presets --> pkg_system_prompt pkg_sandbox_local --> pkg_invariants pkg_sandbox_local --> pkg_llm pkg_sandbox_local --> pkg_sandbox @@ -657,11 +649,6 @@ flowchart TD pkg_llm_pi_ai --> pkg_llm pkg_llm_pi_ai --> pkg_settings pkg_llm_pi_ai --> pkg_timeout - pkg_plugin_package_inventory_deepseek --> pkg_agent - pkg_plugin_package_inventory_deepseek --> pkg_agent_presets - pkg_plugin_package_inventory_deepseek --> pkg_deepseek_llm_api_extensions - pkg_plugin_package_inventory_deepseek --> pkg_invariants - pkg_plugin_package_inventory_deepseek --> pkg_session pkg_tools --> pkg_agent pkg_tools --> pkg_code_runtime pkg_tools --> pkg_invariants @@ -712,8 +699,6 @@ flowchart TD pkg_command_feedback --> pkg_invariants pkg_command_feedback --> pkg_session pkg_command_feedback --> pkg_session_telemetry - pkg_host_apiproxy --> pkg_agent_presets - pkg_host_apiproxy --> pkg_invariants pkg_permission_presets --> pkg_commands pkg_permission_presets --> pkg_invariants pkg_permission_presets --> pkg_sandbox @@ -814,21 +799,6 @@ flowchart TD pkg_tool_skill --> pkg_llm pkg_tool_skill --> pkg_skill pkg_tool_skill --> pkg_tools - pkg_subagent --> pkg_agent - pkg_subagent --> pkg_agent_presets - pkg_subagent --> pkg_brand - pkg_subagent --> pkg_invariants - pkg_subagent --> pkg_jobs - pkg_subagent --> pkg_llm - pkg_subagent --> pkg_sandbox - pkg_subagent --> pkg_sandbox_policy - pkg_subagent --> pkg_scope - pkg_subagent --> pkg_session - pkg_subagent --> pkg_session_persistence - pkg_subagent --> pkg_session_projection - pkg_subagent --> pkg_session_projection_cache - pkg_subagent --> pkg_tools - pkg_subagent --> pkg_user_approval pkg_tool_web --> pkg_invariants pkg_tool_web --> pkg_llm pkg_tool_web --> pkg_system_prompt @@ -876,10 +846,6 @@ flowchart TD pkg_file_reference_local --> pkg_invariants pkg_file_reference_local --> pkg_system_prompt pkg_file_reference_local --> pkg_tools - pkg_experimental_webworker_runtime --> pkg_client_modules - pkg_experimental_webworker_runtime --> pkg_host_apiproxy - pkg_experimental_webworker_runtime --> pkg_host_webserver - pkg_experimental_webworker_runtime --> pkg_invariants pkg_cordis_host_runner --> pkg_agent pkg_cordis_host_runner --> pkg_brand pkg_cordis_host_runner --> pkg_invariants @@ -919,6 +885,15 @@ flowchart TD pkg_mcp_client --> pkg_subprocess pkg_mcp_client --> pkg_timeout pkg_mcp_client --> pkg_tools + pkg_agent_presets --> pkg_agent + pkg_agent_presets --> pkg_atomic_write + pkg_agent_presets --> pkg_home_paths + pkg_agent_presets --> pkg_invariants + pkg_agent_presets --> pkg_scope + pkg_agent_presets --> pkg_session + pkg_agent_presets --> pkg_settings + pkg_agent_presets --> pkg_system_prompt + pkg_agent_presets --> pkg_tools pkg_schedule --> pkg_agent pkg_schedule --> pkg_brand pkg_schedule --> pkg_invariants @@ -992,6 +967,89 @@ flowchart TD pkg_llm_replay --> pkg_invariants pkg_llm_replay --> pkg_llm pkg_llm_replay --> pkg_session + pkg_tool_workflow --> pkg_agent + pkg_tool_workflow --> pkg_invariants + pkg_tool_workflow --> pkg_llm + pkg_tool_workflow --> pkg_session + pkg_tool_workflow --> pkg_system_prompt + pkg_tool_workflow --> pkg_tools + pkg_tool_workflow --> pkg_workflow + pkg_plugin_package_inventory_deepseek --> pkg_agent + pkg_plugin_package_inventory_deepseek --> pkg_agent_presets + pkg_plugin_package_inventory_deepseek --> pkg_deepseek_llm_api_extensions + pkg_plugin_package_inventory_deepseek --> pkg_invariants + pkg_plugin_package_inventory_deepseek --> pkg_session + pkg_subagent --> pkg_agent + pkg_subagent --> pkg_agent_presets + pkg_subagent --> pkg_brand + pkg_subagent --> pkg_invariants + pkg_subagent --> pkg_jobs + pkg_subagent --> pkg_llm + pkg_subagent --> pkg_sandbox + pkg_subagent --> pkg_sandbox_policy + pkg_subagent --> pkg_scope + pkg_subagent --> pkg_session + pkg_subagent --> pkg_session_persistence + pkg_subagent --> pkg_session_projection + pkg_subagent --> pkg_session_projection_cache + pkg_subagent --> pkg_tools + pkg_subagent --> pkg_user_approval + pkg_session_query --> pkg_brand + pkg_session_query --> pkg_invariants + pkg_session_query --> pkg_llm + pkg_session_query --> pkg_session + pkg_session_query --> pkg_session_persistence + pkg_session_query --> pkg_session_title + pkg_session_query --> pkg_tool_todo + pkg_acp --> pkg_agent + pkg_acp --> pkg_attachment + pkg_acp --> pkg_invariants + pkg_acp --> pkg_llm + pkg_acp --> pkg_mcp_client + pkg_acp --> pkg_session + pkg_acp --> pkg_session_persistence + pkg_acp --> pkg_token_meter + pkg_acp --> pkg_user_approval + pkg_web_app --> pkg_invariants + pkg_web_app --> pkg_shell_env + pkg_web_app --> pkg_system_prompt + pkg_compaction_tool_result_pruner --> pkg_compaction + pkg_compaction_tool_result_pruner --> pkg_invariants + pkg_compaction_tool_result_pruner --> pkg_llm + pkg_compaction_tool_result_pruner --> pkg_session + pkg_compaction_tool_result_pruner --> pkg_token_meter + pkg_tool_cordis --> pkg_agent + pkg_tool_cordis --> pkg_cordis_host_runner + pkg_tool_cordis --> pkg_invariants + pkg_tool_cordis --> pkg_llm + pkg_tool_cordis --> pkg_scope + pkg_tool_cordis --> pkg_session + pkg_tool_cordis --> pkg_system_prompt + pkg_tool_cordis --> pkg_tools + pkg_host_apiproxy --> pkg_agent_presets + pkg_host_apiproxy --> pkg_invariants + pkg_tool_bash --> pkg_agent + pkg_tool_bash --> pkg_invariants + pkg_tool_bash --> pkg_jobs + pkg_tool_bash --> pkg_llm + pkg_tool_bash --> pkg_sandbox + pkg_tool_bash --> pkg_sandbox_policy + pkg_tool_bash --> pkg_shell + pkg_tool_bash --> pkg_shell_env + pkg_tool_bash --> pkg_system_prompt + pkg_tool_bash --> pkg_tools + pkg_tool_bash --> pkg_user_approval + pkg_tool_pwsh --> pkg_agent + pkg_tool_pwsh --> pkg_invariants + pkg_tool_pwsh --> pkg_jobs + pkg_tool_pwsh --> pkg_llm + pkg_tool_pwsh --> pkg_sandbox + pkg_tool_pwsh --> pkg_sandbox_policy + pkg_tool_pwsh --> pkg_shell + pkg_tool_pwsh --> pkg_shell_env + pkg_tool_pwsh --> pkg_system_prompt + pkg_tool_pwsh --> pkg_tools + pkg_tool_pwsh --> pkg_user_approval pkg_webhook --> pkg_agent pkg_webhook --> pkg_agent_default_model pkg_webhook --> pkg_agent_presets @@ -1002,13 +1060,6 @@ flowchart TD pkg_webhook --> pkg_session pkg_webhook --> pkg_session_title pkg_webhook --> pkg_workspace - pkg_tool_workflow --> pkg_agent - pkg_tool_workflow --> pkg_invariants - pkg_tool_workflow --> pkg_llm - pkg_tool_workflow --> pkg_session - pkg_tool_workflow --> pkg_system_prompt - pkg_tool_workflow --> pkg_tools - pkg_tool_workflow --> pkg_workflow pkg_subagent_acp --> pkg_agent pkg_subagent_acp --> pkg_invariants pkg_subagent_acp --> pkg_llm @@ -1039,6 +1090,9 @@ flowchart TD pkg_tool_subagent --> pkg_invariants pkg_tool_subagent --> pkg_jobs pkg_tool_subagent --> pkg_llm + pkg_tool_subagent --> pkg_scope + pkg_tool_subagent --> pkg_session + pkg_tool_subagent --> pkg_settings pkg_tool_subagent --> pkg_subagent pkg_tool_subagent --> pkg_system_prompt pkg_tool_subagent --> pkg_tools @@ -1060,107 +1114,6 @@ flowchart TD pkg_hooks_claude_code --> pkg_session_persistence pkg_hooks_claude_code --> pkg_subagent pkg_hooks_claude_code --> pkg_tools - pkg_session_query --> pkg_brand - pkg_session_query --> pkg_invariants - pkg_session_query --> pkg_llm - pkg_session_query --> pkg_session - pkg_session_query --> pkg_session_persistence - pkg_session_query --> pkg_session_title - pkg_session_query --> pkg_tool_todo - pkg_acp --> pkg_agent - pkg_acp --> pkg_attachment - pkg_acp --> pkg_invariants - pkg_acp --> pkg_llm - pkg_acp --> pkg_mcp_client - pkg_acp --> pkg_session - pkg_acp --> pkg_session_persistence - pkg_acp --> pkg_token_meter - pkg_acp --> pkg_user_approval - pkg_web_app --> pkg_invariants - pkg_web_app --> pkg_shell_env - pkg_web_app --> pkg_system_prompt - pkg_client_connection --> pkg_attachment - pkg_client_connection --> pkg_commands - pkg_client_connection --> pkg_host_apiproxy - pkg_client_connection --> pkg_host_webserver - pkg_client_connection --> pkg_invariants - pkg_client_connection --> pkg_llm - pkg_client_connection --> pkg_session - pkg_client_connection --> pkg_tool_todo - pkg_compaction_tool_result_pruner --> pkg_compaction - pkg_compaction_tool_result_pruner --> pkg_invariants - pkg_compaction_tool_result_pruner --> pkg_llm - pkg_compaction_tool_result_pruner --> pkg_session - pkg_compaction_tool_result_pruner --> pkg_token_meter - pkg_experimental_agent_team --> pkg_agent - pkg_experimental_agent_team --> pkg_brand - pkg_experimental_agent_team --> pkg_invariants - pkg_experimental_agent_team --> pkg_llm - pkg_experimental_agent_team --> pkg_session - pkg_experimental_agent_team --> pkg_session_persistence - pkg_experimental_agent_team --> pkg_subagent - pkg_tool_cordis --> pkg_agent - pkg_tool_cordis --> pkg_cordis_host_runner - pkg_tool_cordis --> pkg_invariants - pkg_tool_cordis --> pkg_llm - pkg_tool_cordis --> pkg_scope - pkg_tool_cordis --> pkg_session - pkg_tool_cordis --> pkg_system_prompt - pkg_tool_cordis --> pkg_tools - pkg_sdk_protocol --> pkg_invariants - pkg_sdk_protocol --> pkg_llm - pkg_sdk_protocol --> pkg_session - pkg_sdk_protocol --> pkg_subagent - pkg_tool_bash --> pkg_agent - pkg_tool_bash --> pkg_invariants - pkg_tool_bash --> pkg_jobs - pkg_tool_bash --> pkg_llm - pkg_tool_bash --> pkg_sandbox - pkg_tool_bash --> pkg_sandbox_policy - pkg_tool_bash --> pkg_shell - pkg_tool_bash --> pkg_shell_env - pkg_tool_bash --> pkg_system_prompt - pkg_tool_bash --> pkg_tools - pkg_tool_bash --> pkg_user_approval - pkg_tool_pwsh --> pkg_agent - pkg_tool_pwsh --> pkg_invariants - pkg_tool_pwsh --> pkg_jobs - pkg_tool_pwsh --> pkg_llm - pkg_tool_pwsh --> pkg_sandbox - pkg_tool_pwsh --> pkg_sandbox_policy - pkg_tool_pwsh --> pkg_shell - pkg_tool_pwsh --> pkg_shell_env - pkg_tool_pwsh --> pkg_system_prompt - pkg_tool_pwsh --> pkg_tools - pkg_tool_pwsh --> pkg_user_approval - pkg_webhook_github --> pkg_credentials - pkg_webhook_github --> pkg_host_webserver - pkg_webhook_github --> pkg_invariants - pkg_webhook_github --> pkg_session - pkg_webhook_github --> pkg_webhook - pkg_tool_ralph --> pkg_agent - pkg_tool_ralph --> pkg_invariants - pkg_tool_ralph --> pkg_llm - pkg_tool_ralph --> pkg_subagent - pkg_tool_ralph --> pkg_system_prompt - pkg_tool_ralph --> pkg_tools - pkg_tool_ralph --> pkg_workflow - pkg_workflow_worker_thread --> pkg_agent - pkg_workflow_worker_thread --> pkg_brand - pkg_workflow_worker_thread --> pkg_invariants - pkg_workflow_worker_thread --> pkg_llm - pkg_workflow_worker_thread --> pkg_session - pkg_workflow_worker_thread --> pkg_subagent - pkg_workflow_worker_thread --> pkg_tools - pkg_workflow_worker_thread --> pkg_workflow - pkg_subagent_fork_in_process --> pkg_agent - pkg_subagent_fork_in_process --> pkg_invariants - pkg_subagent_fork_in_process --> pkg_session - pkg_subagent_fork_in_process --> pkg_subagent - pkg_subagent_fork_in_process --> pkg_subagent_in_process_driver - pkg_subagent_spawn_in_process --> pkg_invariants - pkg_subagent_spawn_in_process --> pkg_subagent - pkg_subagent_spawn_in_process --> pkg_subagent_in_process_driver pkg_session_query_sqlite --> pkg_invariants pkg_session_query_sqlite --> pkg_session pkg_session_query_sqlite --> pkg_session_persistence @@ -1172,11 +1125,14 @@ flowchart TD pkg_tool_session_query --> pkg_system_prompt pkg_tool_session_query --> pkg_timeout pkg_tool_session_query --> pkg_tools - pkg_api_gateway --> pkg_brand - pkg_api_gateway --> pkg_client_connection - pkg_api_gateway --> pkg_host_webserver - pkg_api_gateway --> pkg_invariants - pkg_api_gateway --> pkg_typert_registry + pkg_client_connection --> pkg_attachment + pkg_client_connection --> pkg_commands + pkg_client_connection --> pkg_host_apiproxy + pkg_client_connection --> pkg_host_webserver + pkg_client_connection --> pkg_invariants + pkg_client_connection --> pkg_llm + pkg_client_connection --> pkg_session + pkg_client_connection --> pkg_tool_todo pkg_compaction_basic --> pkg_agent pkg_compaction_basic --> pkg_commands pkg_compaction_basic --> pkg_compaction @@ -1215,6 +1171,54 @@ flowchart TD pkg_agent_spine_demo --> pkg_tool_jobs pkg_agent_spine_demo --> pkg_tool_skill pkg_agent_spine_demo --> pkg_tools + pkg_experimental_agent_team --> pkg_agent + pkg_experimental_agent_team --> pkg_brand + pkg_experimental_agent_team --> pkg_invariants + pkg_experimental_agent_team --> pkg_llm + pkg_experimental_agent_team --> pkg_session + pkg_experimental_agent_team --> pkg_session_persistence + pkg_experimental_agent_team --> pkg_subagent + pkg_experimental_webworker_runtime --> pkg_client_modules + pkg_experimental_webworker_runtime --> pkg_host_apiproxy + pkg_experimental_webworker_runtime --> pkg_host_webserver + pkg_experimental_webworker_runtime --> pkg_invariants + pkg_sdk_protocol --> pkg_invariants + pkg_sdk_protocol --> pkg_llm + pkg_sdk_protocol --> pkg_session + pkg_sdk_protocol --> pkg_subagent + pkg_webhook_github --> pkg_credentials + pkg_webhook_github --> pkg_host_webserver + pkg_webhook_github --> pkg_invariants + pkg_webhook_github --> pkg_session + pkg_webhook_github --> pkg_webhook + pkg_tool_ralph --> pkg_agent + pkg_tool_ralph --> pkg_invariants + pkg_tool_ralph --> pkg_llm + pkg_tool_ralph --> pkg_subagent + pkg_tool_ralph --> pkg_system_prompt + pkg_tool_ralph --> pkg_tools + pkg_tool_ralph --> pkg_workflow + pkg_workflow_worker_thread --> pkg_agent + pkg_workflow_worker_thread --> pkg_brand + pkg_workflow_worker_thread --> pkg_invariants + pkg_workflow_worker_thread --> pkg_llm + pkg_workflow_worker_thread --> pkg_session + pkg_workflow_worker_thread --> pkg_subagent + pkg_workflow_worker_thread --> pkg_tools + pkg_workflow_worker_thread --> pkg_workflow + pkg_subagent_fork_in_process --> pkg_agent + pkg_subagent_fork_in_process --> pkg_invariants + pkg_subagent_fork_in_process --> pkg_session + pkg_subagent_fork_in_process --> pkg_subagent + pkg_subagent_fork_in_process --> pkg_subagent_in_process_driver + pkg_subagent_spawn_in_process --> pkg_invariants + pkg_subagent_spawn_in_process --> pkg_subagent + pkg_subagent_spawn_in_process --> pkg_subagent_in_process_driver + pkg_api_gateway --> pkg_brand + pkg_api_gateway --> pkg_client_connection + pkg_api_gateway --> pkg_host_webserver + pkg_api_gateway --> pkg_invariants + pkg_api_gateway --> pkg_typert_registry pkg_experimental_tool_agent_team --> pkg_agent pkg_experimental_tool_agent_team --> pkg_experimental_agent_team pkg_experimental_tool_agent_team --> pkg_invariants @@ -1642,7 +1646,7 @@ flowchart TD pkg_client_ui_cordis --> pkg_invariants ``` -| Package | Group | Depends on | +| 包 | 分组 | 依赖 | | --- | --- | --- | | [`invariants`](../packages/runtime-diagnostics/invariants) | `runtime-diagnostics` | — | | [`atomic-write`](../packages/util/atomic-write) | `util` | [`invariants`](../packages/runtime-diagnostics/invariants) | @@ -1733,7 +1737,6 @@ flowchart TD | [`user-approval`](../packages/interaction/user-approval) | `interaction` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt) | | [`user-questions`](../packages/interaction/user-questions) | `interaction` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope) | | [`jobs`](../packages/jobs/jobs) | `jobs` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session) | -| [`agent-presets`](../packages/preset/agent-presets) | `preset` | [`agent`](../packages/core/agent), [`atomic-write`](../packages/util/atomic-write), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`settings`](../packages/settings/settings), [`system-prompt`](../packages/core/system-prompt) | | [`sandbox-local`](../packages/sandbox/sandbox-local) | `sandbox` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`session`](../packages/core/session) | | [`sandbox-policy`](../packages/sandbox/sandbox-policy) | `sandbox` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`sandbox`](../packages/sandbox/sandbox), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt) | | [`session-persistence-jsonl`](../packages/session/session-persistence-jsonl) | `session` | [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence) | @@ -1749,7 +1752,6 @@ flowchart TD | [`workspace`](../packages/workspace/workspace) | `workspace` | [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`storage`](../packages/storage/storage), [`storage-domain`](../packages/storage/storage-domain) | | [`llm-deepseek`](../packages/llm/llm-deepseek) | `llm` | [`anonymous-user-id`](../packages/identity/anonymous-user-id), [`atomic-write`](../packages/util/atomic-write), [`attachment`](../packages/attachment/attachment), [`brand`](../packages/util/brand), [`credentials`](../packages/credentials/credentials), [`deepseek-llm-api-extensions`](../packages/llm/deepseek-llm-api-extensions), [`fs`](../packages/fs/fs), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`launch-environment`](../packages/util/launch-environment), [`llm`](../packages/llm/llm), [`settings`](../packages/settings/settings), [`timeout`](../packages/util/timeout) | | [`llm-pi-ai`](../packages/llm/llm-pi-ai) | `llm` | [`attachment`](../packages/attachment/attachment), [`authorization`](../packages/credentials/authorization), [`credentials`](../packages/credentials/credentials), [`fs`](../packages/fs/fs), [`invariants`](../packages/runtime-diagnostics/invariants), [`launch-environment`](../packages/util/launch-environment), [`llm`](../packages/llm/llm), [`settings`](../packages/settings/settings), [`timeout`](../packages/util/timeout) | -| [`plugin-package-inventory-deepseek`](../packages/llm/plugin-package-inventory-deepseek) | `llm` | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`deepseek-llm-api-extensions`](../packages/llm/deepseek-llm-api-extensions), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session) | | [`tools`](../packages/core/tools) | `core` | [`agent`](../packages/core/agent), [`code-runtime`](../packages/code-runtime/code-runtime), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`user-approval`](../packages/interaction/user-approval) | | [`command-goal`](../packages/goal/command-goal) | `goal` | [`commands`](../packages/interaction/commands), [`goal`](../packages/goal/goal), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm) | | [`goal-round-driver`](../packages/goal/goal-round-driver) | `goal` | [`agent`](../packages/core/agent), [`goal`](../packages/goal/goal), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session) | @@ -1762,7 +1764,6 @@ flowchart TD | [`tmux-context`](../packages/context/tmux-context) | `context` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`shell`](../packages/shell/shell) | | [`fs-e2b`](../packages/e2b/fs-e2b) | `e2b` | [`e2b`](../packages/e2b/e2b), [`fs`](../packages/fs/fs), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`command-feedback`](../packages/feedback/command-feedback) | `feedback` | [`anonymous-user-id`](../packages/identity/anonymous-user-id), [`commands`](../packages/interaction/commands), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-telemetry`](../packages/session/session-telemetry) | -| [`host-apiproxy`](../packages/host/apiproxy) | `host` | [`agent-presets`](../packages/preset/agent-presets), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`permission-presets`](../packages/interaction/permission-presets) | `interaction` | [`commands`](../packages/interaction/commands), [`invariants`](../packages/runtime-diagnostics/invariants), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`session`](../packages/core/session), [`session-projection`](../packages/session/session-projection), [`settings`](../packages/settings/settings), [`shell`](../packages/shell/shell), [`user-approval`](../packages/interaction/user-approval) | | [`jobs-local`](../packages/jobs/jobs-local) | `jobs` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`scope`](../packages/core/scope), [`timeout`](../packages/util/timeout) | | [`lsp-stdio`](../packages/lsp/lsp-stdio) | `lsp` | [`brand`](../packages/util/brand), [`fs`](../packages/fs/fs), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`lsp`](../packages/lsp/lsp), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout) | @@ -1779,7 +1780,6 @@ flowchart TD | [`tool-fs-search`](../packages/fs/tool-fs-search) | `fs` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`session`](../packages/core/session), [`spill`](../packages/spill/spill), [`subprocess`](../packages/subprocess/subprocess), [`system-prompt`](../packages/core/system-prompt), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | | [`tool-str-replace-editor`](../packages/fs/tool-str-replace-editor) | `fs` | [`fs`](../packages/fs/fs), [`invariants`](../packages/runtime-diagnostics/invariants), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`tools`](../packages/core/tools) | | [`tool-skill`](../packages/skill/tool-skill) | `skill` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`skill`](../packages/skill/skill), [`tools`](../packages/core/tools) | -| [`subagent`](../packages/subagent/subagent) | `subagent` | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-projection`](../packages/session/session-projection), [`session-projection-cache`](../packages/session/session-projection-cache), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | | [`tool-web`](../packages/web/tool-web) | `web` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`web`](../packages/web/web) | | [`spill-policy`](../packages/spill/spill-policy) | `spill` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`session`](../packages/core/session), [`spill`](../packages/spill/spill), [`tools`](../packages/core/tools) | | [`tool-todo`](../packages/todo/tool-todo) | `todo` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-projection`](../packages/session/session-projection), [`tools`](../packages/core/tools) | @@ -1788,7 +1788,6 @@ flowchart TD | [`command-compact`](../packages/compaction/command-compact) | `compaction` | [`commands`](../packages/interaction/commands), [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`agent-instructions`](../packages/context/agent-instructions) | `context` | [`agent`](../packages/core/agent), [`fs`](../packages/fs/fs), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`tools`](../packages/core/tools) | | [`file-reference-local`](../packages/context/file-reference-local) | `context` | [`agent`](../packages/core/agent), [`file-reference`](../packages/context/file-reference), [`invariants`](../packages/runtime-diagnostics/invariants), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | -| [`experimental-webworker-runtime`](../packages/experimental/webworker-runtime) | `experimental` | [`client-modules`](../packages/client/modules), [`host-apiproxy`](../packages/host/apiproxy), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`cordis-host-runner`](../packages/extensions/cordis-host-runner) | `extensions` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`tools`](../packages/core/tools), [`typert-protocol`](../packages/typert/protocol) | | [`repeat-tool-reminder`](../packages/guard/repeat-tool-reminder) | `guard` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`tools`](../packages/core/tools) | | [`tool-call-timeout-policy`](../packages/guard/timeout-policy) | `guard` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | @@ -1796,6 +1795,7 @@ flowchart TD | [`tool-jobs`](../packages/jobs/tool-jobs) | `jobs` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`tool-lsp`](../packages/lsp/tool-lsp) | `lsp` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`lsp`](../packages/lsp/lsp), [`system-prompt`](../packages/core/system-prompt), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | | [`mcp-client`](../packages/mcp/mcp-client) | `mcp` | [`attachment`](../packages/attachment/attachment), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | +| [`agent-presets`](../packages/preset/agent-presets) | `preset` | [`agent`](../packages/core/agent), [`atomic-write`](../packages/util/atomic-write), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`settings`](../packages/settings/settings), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`schedule`](../packages/schedule/schedule) | `schedule` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`tools`](../packages/core/tools) | | [`session-checkpoint-policy`](../packages/session/session-checkpoint-policy) | `session` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`tools`](../packages/core/tools) | | [`session-telemetry-otel`](../packages/session/session-telemetry-otel) | `session` | [`anonymous-user-id`](../packages/identity/anonymous-user-id), [`command-feedback`](../packages/feedback/command-feedback), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-telemetry`](../packages/session/session-telemetry) | @@ -1809,37 +1809,41 @@ flowchart TD | [`tool-terminal`](../packages/terminal/tool-terminal) | `terminal` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`system-prompt`](../packages/core/system-prompt), [`terminal`](../packages/terminal/terminal), [`tools`](../packages/core/tools) | | [`agent-loop-testkit`](../packages/test-support/agent-loop-testkit) | `test-support` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`llm-replay`](../packages/test-support/llm-replay) | `test-support` | [`compaction`](../packages/compaction/compaction), [`deepseek-llm-api-extensions`](../packages/llm/deepseek-llm-api-extensions), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session) | -| [`webhook`](../packages/webhook/webhook) | `webhook` | [`agent`](../packages/core/agent), [`agent-default-model`](../packages/core/agent-default-model), [`agent-presets`](../packages/preset/agent-presets), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`permission-presets`](../packages/interaction/permission-presets), [`session`](../packages/core/session), [`session-title`](../packages/session/session-title), [`workspace`](../packages/workspace/workspace) | | [`tool-workflow`](../packages/workflow/tool-workflow) | `workflow` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`workflow`](../packages/workflow/workflow) | +| [`plugin-package-inventory-deepseek`](../packages/llm/plugin-package-inventory-deepseek) | `llm` | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`deepseek-llm-api-extensions`](../packages/llm/deepseek-llm-api-extensions), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session) | +| [`subagent`](../packages/subagent/subagent) | `subagent` | [`agent`](../packages/core/agent), [`agent-presets`](../packages/preset/agent-presets), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-projection`](../packages/session/session-projection), [`session-projection-cache`](../packages/session/session-projection-cache), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | +| [`session-query`](../packages/session-query/session-query) | `session-query` | [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-title`](../packages/session/session-title), [`tool-todo`](../packages/todo/tool-todo) | +| [`acp`](../packages/acp/acp) | `acp` | [`agent`](../packages/core/agent), [`attachment`](../packages/attachment/attachment), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`mcp-client`](../packages/mcp/mcp-client), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`token-meter`](../packages/llm/token-meter), [`user-approval`](../packages/interaction/user-approval) | +| [`web-app`](../packages/bundle/web-app) | `bundle` | [`invariants`](../packages/runtime-diagnostics/invariants), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt) | +| [`compaction-tool-result-pruner`](../packages/compaction/compaction-tool-result-pruner) | `compaction` | [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`token-meter`](../packages/llm/token-meter) | +| [`tool-cordis`](../packages/extensions/tool-cordis) | `extensions` | [`agent`](../packages/core/agent), [`cordis-host-runner`](../packages/extensions/cordis-host-runner), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | +| [`host-apiproxy`](../packages/host/apiproxy) | `host` | [`agent-presets`](../packages/preset/agent-presets), [`invariants`](../packages/runtime-diagnostics/invariants) | +| [`tool-bash`](../packages/shell/tool-bash) | `shell` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`shell`](../packages/shell/shell), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | +| [`tool-pwsh`](../packages/shell/tool-pwsh) | `shell` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`shell`](../packages/shell/shell), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | +| [`webhook`](../packages/webhook/webhook) | `webhook` | [`agent`](../packages/core/agent), [`agent-default-model`](../packages/core/agent-default-model), [`agent-presets`](../packages/preset/agent-presets), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`permission-presets`](../packages/interaction/permission-presets), [`session`](../packages/core/session), [`session-title`](../packages/session/session-title), [`workspace`](../packages/workspace/workspace) | | [`subagent-acp`](../packages/subagent/subagent-acp) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout) | | [`subagent-claude-code`](../packages/subagent/subagent-claude-code) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout) | | [`subagent-codex`](../packages/subagent/subagent-codex) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`subprocess`](../packages/subprocess/subprocess), [`timeout`](../packages/util/timeout) | | [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | -| [`tool-subagent`](../packages/subagent/tool-subagent) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | +| [`tool-subagent`](../packages/subagent/tool-subagent) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`settings`](../packages/settings/settings), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`tool-subagent-control`](../packages/subagent/tool-subagent-control) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`tools`](../packages/core/tools) | | [`tool-subagent-report`](../packages/subagent/tool-subagent-report) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`hooks-claude-code`](../packages/hooks/hooks-claude-code) | `hooks` | [`agent`](../packages/core/agent), [`hook-protocol`](../packages/hooks/hook-protocol), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`subagent`](../packages/subagent/subagent), [`tools`](../packages/core/tools) | -| [`session-query`](../packages/session-query/session-query) | `session-query` | [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-title`](../packages/session/session-title), [`tool-todo`](../packages/todo/tool-todo) | -| [`acp`](../packages/acp/acp) | `acp` | [`agent`](../packages/core/agent), [`attachment`](../packages/attachment/attachment), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`mcp-client`](../packages/mcp/mcp-client), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`token-meter`](../packages/llm/token-meter), [`user-approval`](../packages/interaction/user-approval) | -| [`web-app`](../packages/bundle/web-app) | `bundle` | [`invariants`](../packages/runtime-diagnostics/invariants), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt) | +| [`session-query-sqlite`](../packages/session-query/session-query-sqlite) | `session-query` | [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-query`](../packages/session-query/session-query) | +| [`tool-session-query`](../packages/session-query/tool-session-query) | `session-query` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-query`](../packages/session-query/session-query), [`system-prompt`](../packages/core/system-prompt), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | | [`client-connection`](../packages/client/connection) | `client` | [`attachment`](../packages/attachment/attachment), [`commands`](../packages/interaction/commands), [`host-apiproxy`](../packages/host/apiproxy), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`tool-todo`](../packages/todo/tool-todo) | -| [`compaction-tool-result-pruner`](../packages/compaction/compaction-tool-result-pruner) | `compaction` | [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`token-meter`](../packages/llm/token-meter) | +| [`compaction-basic`](../packages/compaction/compaction-basic) | `compaction` | [`agent`](../packages/core/agent), [`commands`](../packages/interaction/commands), [`compaction`](../packages/compaction/compaction), [`compaction-tool-result-pruner`](../packages/compaction/compaction-tool-result-pruner), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`token-meter`](../packages/llm/token-meter) | +| [`session-reference`](../packages/context/session-reference) | `context` | [`agent`](../packages/core/agent), [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`session`](../packages/core/session), [`session-query`](../packages/session-query/session-query), [`typert-protocol`](../packages/typert/protocol) | +| [`agent-spine-demo`](../packages/examples/agent-spine-demo) | `examples` | [`agent`](../packages/core/agent), [`agent-instructions`](../packages/context/agent-instructions), [`agent-loop`](../packages/core/agent-loop), [`goal`](../packages/goal/goal), [`goal-round-driver`](../packages/goal/goal-round-driver), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs-local`](../packages/jobs/jobs-local), [`llm`](../packages/llm/llm), [`llm-retry`](../packages/llm/llm-retry), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`session-title`](../packages/session/session-title), [`shell-env`](../packages/shell/shell-env), [`skill`](../packages/skill/skill), [`skill-filesystem`](../packages/skill/skill-filesystem), [`system-prompt`](../packages/core/system-prompt), [`tool-bash`](../packages/shell/tool-bash), [`tool-goal`](../packages/goal/tool-goal), [`tool-jobs`](../packages/jobs/tool-jobs), [`tool-skill`](../packages/skill/tool-skill), [`tools`](../packages/core/tools) | | [`experimental-agent-team`](../packages/experimental/agent-team) | `experimental` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`subagent`](../packages/subagent/subagent) | -| [`tool-cordis`](../packages/extensions/tool-cordis) | `extensions` | [`agent`](../packages/core/agent), [`cordis-host-runner`](../packages/extensions/cordis-host-runner), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | +| [`experimental-webworker-runtime`](../packages/experimental/webworker-runtime) | `experimental` | [`client-modules`](../packages/client/modules), [`host-apiproxy`](../packages/host/apiproxy), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants) | | [`sdk-protocol`](../packages/sdk/protocol) | `sdk` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent) | -| [`tool-bash`](../packages/shell/tool-bash) | `shell` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`shell`](../packages/shell/shell), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | -| [`tool-pwsh`](../packages/shell/tool-pwsh) | `shell` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs`](../packages/jobs/jobs), [`llm`](../packages/llm/llm), [`sandbox`](../packages/sandbox/sandbox), [`sandbox-policy`](../packages/sandbox/sandbox-policy), [`shell`](../packages/shell/shell), [`shell-env`](../packages/shell/shell-env), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`user-approval`](../packages/interaction/user-approval) | | [`webhook-github`](../packages/webhook/webhook-github) | `webhook` | [`credentials`](../packages/credentials/credentials), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`webhook`](../packages/webhook/webhook) | | [`tool-ralph`](../packages/workflow/tool-ralph) | `workflow` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`subagent`](../packages/subagent/subagent), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools), [`workflow`](../packages/workflow/workflow) | | [`workflow-worker-thread`](../packages/workflow/workflow-worker-thread) | `workflow` | [`agent`](../packages/core/agent), [`brand`](../packages/util/brand), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`tools`](../packages/core/tools), [`workflow`](../packages/workflow/workflow) | | [`subagent-fork-in-process`](../packages/subagent/subagent-fork-in-process) | `subagent` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent), [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver) | | [`subagent-spawn-in-process`](../packages/subagent/subagent-spawn-in-process) | `subagent` | [`invariants`](../packages/runtime-diagnostics/invariants), [`subagent`](../packages/subagent/subagent), [`subagent-in-process-driver`](../packages/subagent/subagent-in-process-driver) | -| [`session-query-sqlite`](../packages/session-query/session-query-sqlite) | `session-query` | [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`session-persistence`](../packages/session/session-persistence), [`session-query`](../packages/session-query/session-query) | -| [`tool-session-query`](../packages/session-query/tool-session-query) | `session-query` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`session-query`](../packages/session-query/session-query), [`system-prompt`](../packages/core/system-prompt), [`timeout`](../packages/util/timeout), [`tools`](../packages/core/tools) | | [`api-gateway`](../packages/api/gateway) | `api` | [`brand`](../packages/util/brand), [`client-connection`](../packages/client/connection), [`host-webserver`](../packages/host/webserver), [`invariants`](../packages/runtime-diagnostics/invariants), [`typert-registry`](../packages/typert/registry) | -| [`compaction-basic`](../packages/compaction/compaction-basic) | `compaction` | [`agent`](../packages/core/agent), [`commands`](../packages/interaction/commands), [`compaction`](../packages/compaction/compaction), [`compaction-tool-result-pruner`](../packages/compaction/compaction-tool-result-pruner), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`session`](../packages/core/session), [`token-meter`](../packages/llm/token-meter) | -| [`session-reference`](../packages/context/session-reference) | `context` | [`agent`](../packages/core/agent), [`compaction`](../packages/compaction/compaction), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`output-retention`](../packages/util/output-retention), [`session`](../packages/core/session), [`session-query`](../packages/session-query/session-query), [`typert-protocol`](../packages/typert/protocol) | -| [`agent-spine-demo`](../packages/examples/agent-spine-demo) | `examples` | [`agent`](../packages/core/agent), [`agent-instructions`](../packages/context/agent-instructions), [`agent-loop`](../packages/core/agent-loop), [`goal`](../packages/goal/goal), [`goal-round-driver`](../packages/goal/goal-round-driver), [`home-paths`](../packages/util/home-paths), [`invariants`](../packages/runtime-diagnostics/invariants), [`jobs-local`](../packages/jobs/jobs-local), [`llm`](../packages/llm/llm), [`llm-retry`](../packages/llm/llm-retry), [`scope`](../packages/core/scope), [`session`](../packages/core/session), [`session-title`](../packages/session/session-title), [`shell-env`](../packages/shell/shell-env), [`skill`](../packages/skill/skill), [`skill-filesystem`](../packages/skill/skill-filesystem), [`system-prompt`](../packages/core/system-prompt), [`tool-bash`](../packages/shell/tool-bash), [`tool-goal`](../packages/goal/tool-goal), [`tool-jobs`](../packages/jobs/tool-jobs), [`tool-skill`](../packages/skill/tool-skill), [`tools`](../packages/core/tools) | | [`experimental-tool-agent-team`](../packages/experimental/tool-agent-team) | `experimental` | [`agent`](../packages/core/agent), [`experimental-agent-team`](../packages/experimental/agent-team), [`invariants`](../packages/runtime-diagnostics/invariants), [`session`](../packages/core/session), [`system-prompt`](../packages/core/system-prompt), [`tools`](../packages/core/tools) | | [`sdk-client`](../packages/sdk/client) | `sdk` | [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`sdk-protocol`](../packages/sdk/protocol), [`session`](../packages/core/session) | | [`sdk-jsonrpc-server`](../packages/sdk/server) | `sdk` | [`agent`](../packages/core/agent), [`invariants`](../packages/runtime-diagnostics/invariants), [`llm`](../packages/llm/llm), [`llm-deepseek`](../packages/llm/llm-deepseek), [`scope`](../packages/core/scope), [`sdk-protocol`](../packages/sdk/protocol), [`session`](../packages/core/session), [`subagent`](../packages/subagent/subagent) | diff --git a/docs/persistence-catalog.i18n.yaml b/docs/persistence-catalog.i18n.yaml index 06c60504dc..aad7db87f6 100644 --- a/docs/persistence-catalog.i18n.yaml +++ b/docs/persistence-catalog.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/persistence-catalog.md -persistence-catalog.md: 5155968af0886a389d0b01d9332af927cff95a55 -persistence-catalog.zh.md: abd4ae767a5cfca45be76b54d392815244070869 +persistence-catalog.md: dd2124520e43e590fc3506b23c533b3e482132cd +persistence-catalog.zh.md: 48c0867f37fc87ce5d29ce04b0b6970fca83648c diff --git a/docs/persistence-catalog.md b/docs/persistence-catalog.md index 5155968af0..dd2124520e 100644 --- a/docs/persistence-catalog.md +++ b/docs/persistence-catalog.md @@ -726,7 +726,23 @@ Source: [`packages/core/session/src/types.ts:237`](../packages/core/session/src/ 'subagent/descriptor': SubagentDescriptorData ``` -Source: [`packages/subagent/subagent/src/descriptor.ts:37`](../packages/subagent/subagent/src/descriptor.ts) +Source: [`packages/subagent/subagent/src/descriptor.ts:38`](../packages/subagent/subagent/src/descriptor.ts) + + + +#### `subagent/model-selection-enabled` — log-only + +```ts persistence-catalog +/** + * Records that this session's delegation tool exposes child provider, + * model, and reasoning-effort selection. Appended before the first model + * request; absence means the fixed-route definition. Log-only: it carries + * no `surfaceOp` and never enters model history. + */ +'subagent/model-selection-enabled': Record +``` + +Source: [`packages/subagent/tool-subagent/src/model-selection-state.ts:13`](../packages/subagent/tool-subagent/src/model-selection-state.ts) ### `team/*` diff --git a/docs/persistence-catalog.zh.md b/docs/persistence-catalog.zh.md index abd4ae767a..48c0867f37 100644 --- a/docs/persistence-catalog.zh.md +++ b/docs/persistence-catalog.zh.md @@ -728,7 +728,23 @@ export type SessionEvent = { 'subagent/descriptor': SubagentDescriptorData ``` -来源：[`packages/subagent/subagent/src/descriptor.ts:37`](../packages/subagent/subagent/src/descriptor.ts) +来源：[`packages/subagent/subagent/src/descriptor.ts:38`](../packages/subagent/subagent/src/descriptor.ts) + + + +#### `subagent/model-selection-enabled` — log-only + +```ts persistence-catalog +/** + * Records that this session's delegation tool exposes child provider, + * model, and reasoning-effort selection. Appended before the first model + * request; absence means the fixed-route definition. Log-only: it carries + * no `surfaceOp` and never enters model history. + */ +'subagent/model-selection-enabled': Record +``` + +来源：[`packages/subagent/tool-subagent/src/model-selection-state.ts:13`](../packages/subagent/tool-subagent/src/model-selection-state.ts) ### `team/*` diff --git a/docs/subsystems/core.i18n.yaml b/docs/subsystems/core.i18n.yaml index 9174e27772..fb5dd11fbb 100644 --- a/docs/subsystems/core.i18n.yaml +++ b/docs/subsystems/core.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/subsystems/core.md -core.md: 3417560539f41009a290f4c31b255f338624d56d -core.zh.md: 75f3c0ca16e1235cb2155f6e54e86e414e61b498 +core.md: c53bb94fa5918c3a91ee9aedbb2416d0b101b240 +core.zh.md: 93ee45fb88cc100eb77673f2b70e86483c7ed29f diff --git a/docs/subsystems/core.md b/docs/subsystems/core.md index 3417560539..c53bb94fa5 100644 --- a/docs/subsystems/core.md +++ b/docs/subsystems/core.md @@ -161,12 +161,14 @@ interface AgentOptions { provider?: string /** Model id interpreted by the selected provider adapter. */ model?: string + /** Adapter-owned reasoning effort for the selected provider/model route. */ + reasoningEffort?: ReasoningEffortId /** Maximum output tokens for each conversation-model request. */ maxTokens?: number } ``` -Dispatch requires `provider` and `model` after `agent/request`. When present, `maxTokens` must be a positive safe integer and caps every conversation-model request; omission allows the exact-model adapter default to materialize before the request header, or otherwise leaves provider behavior unchanged. An agent-scoped `deployment:persona` prompt section may shadow the global default persona. +Dispatch requires `provider` and `model` after `agent/request`. An explicit `reasoningEffort` seeds the first request on that route; exact-model resolution validates it, while omission allows the adapter default to materialize. When present, `maxTokens` must be a positive safe integer and caps every conversation-model request; omission allows the exact-model adapter default to materialize before the request header, or otherwise leaves provider behavior unchanged. An agent-scoped `deployment:persona` prompt section may shadow the global default persona. The inbox is the delivery vocabulary — two ordered pending-message lists the agent owns as a durable projection: @@ -524,7 +526,9 @@ serviceFor(agent: { ctx: Context }, name: K): * state to restore. The re-link runs through the binding this roster kept * from the agent's mount — dsh-scope's only re-link authority. An agent * that never composed one has nothing to re-link: the switch is then the - * agent's first bind, exactly a mount. + * agent's first bind, exactly a mount. A committed re-link emits + * `tools/change` because changing the parent scope changes the Agent's + * resolved tool set without adding or removing registry entries. * @param agentCtx - the agent's scope context. * @param id - the preset to compose the agent from instead. * @returns the preset now installed. diff --git a/docs/subsystems/core.zh.md b/docs/subsystems/core.zh.md index 75f3c0ca16..93ee45fb88 100644 --- a/docs/subsystems/core.zh.md +++ b/docs/subsystems/core.zh.md @@ -165,12 +165,14 @@ interface AgentOptions { provider?: string /** Model id interpreted by the selected provider adapter. */ model?: string + /** Adapter-owned reasoning effort for the selected provider/model route. */ + reasoningEffort?: ReasoningEffortId /** Maximum output tokens for each conversation-model request. */ maxTokens?: number } ``` -在 `agent/request` 之后，分发要求 `provider` 与 `model` 都存在。提供 `maxTokens` 时，它必须是正安全整数，并限制每次对话模型请求的输出；省略时，系统会在写入请求 header 前填入确切模型的适配器默认值，否则提供方行为保持不变。agent 作用域的 `deployment:persona` 提示词段落可以遮蔽全局默认 persona。 +在 `agent/request` 之后，分发要求 `provider` 与 `model` 都存在。显式 `reasoningEffort` 会为该路由的首次请求提供初始值；确切模型解析会校验该值，省略时则允许填入适配器默认值。提供 `maxTokens` 时，它必须是正安全整数，并限制每次对话模型请求的输出；省略时，系统会在写入请求 header 前填入确切模型的适配器默认值，否则提供方行为保持不变。agent 作用域的 `deployment:persona` 提示词段落可以遮蔽全局默认 persona。 inbox 即投递词汇——agent 以持久投影形式拥有的两条有序待处理消息列表： @@ -534,7 +536,9 @@ serviceFor(agent: { ctx: Context }, name: K): * state to restore. The re-link runs through the binding this roster kept * from the agent's mount — dsh-scope's only re-link authority. An agent * that never composed one has nothing to re-link: the switch is then the - * agent's first bind, exactly a mount. + * agent's first bind, exactly a mount. A committed re-link emits + * `tools/change` because changing the parent scope changes the Agent's + * resolved tool set without adding or removing registry entries. * @param agentCtx - the agent's scope context. * @param id - the preset to compose the agent from instead. * @returns the preset now installed. diff --git a/docs/subsystems/subagent.i18n.yaml b/docs/subsystems/subagent.i18n.yaml index ac6150111a..cd3f9d99c9 100644 --- a/docs/subsystems/subagent.i18n.yaml +++ b/docs/subsystems/subagent.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/subsystems/subagent.md -subagent.md: 7a0ba28afbbec88465415b3411946a0afa0da918 -subagent.zh.md: f397c19a0dc79c0c5a94e617e2013b68faa4d2e9 +subagent.md: 63edef0a4b8d5368ea9d6d82f0ea3ef99ea0bbad +subagent.zh.md: 21b0dfd21dbee5e1d37558d6c02fe7949126d9a9 diff --git a/docs/subsystems/subagent.md b/docs/subsystems/subagent.md index 7a0ba28afb..63edef0a4b 100644 --- a/docs/subsystems/subagent.md +++ b/docs/subsystems/subagent.md @@ -25,6 +25,7 @@ A provider advertises its **start-time** features on a static descriptor the ser * to `maxDepth`; the other names match. */ interface SubagentCapabilities { + readonly agentOptions: boolean readonly outputSchema: boolean readonly depthLimit: boolean readonly toolFilter: boolean @@ -34,7 +35,7 @@ interface SubagentCapabilities { ## The one-shot start request -The tool layer builds this request from the model input and its own config; the service validates it against the named provider before `start`. Required `parent` supplies the session cwd, lineage, and delegation depth. Optional output schema, depth, tool filter, and persona require matching capability flags. Unsupported schemas fail at start; in-process backends scope filters and personas to child creation and implement the supported object-rooted schema with a forced capture tool. +The tool layer builds this request from the model input and its own config; the service validates it against the named provider before `start`. Required `parent` supplies the session cwd, lineage, and delegation depth. Optional Agent provider, model, reasoning-effort, and token overrides, output schema, depth, tool filter, and persona require matching capability flags. In-process backends merge `agentOptions` over the parent Agent's options, scope filters and personas to child creation, and implement the supported object-rooted schema with a forced capture tool. Current out-of-process providers reject `agentOptions` before starting their transport. ```ts type-equiv /** @@ -63,6 +64,12 @@ interface SubagentStartRequest { * remaining turn work when it fires afterward. */ readonly signal: AbortSignal + /** + * Optional host-Agent provider, model, reasoning-effort, and output-token + * overrides. Requires {@link SubagentCapabilities.agentOptions}; in-process + * providers merge them over the parent Agent's options when they create the + * child. + */ readonly agentOptions?: AgentOptions /** * Object-rooted JSON Schema within `assertObjectJsonSchema`'s enforced subset. Start rejects @@ -280,7 +287,7 @@ interface ContinuableCreateSpec { } ``` -The descriptor (`SubagentDescriptorData` in [descriptor.ts](../../packages/subagent/subagent/src/descriptor.ts)) is a mode-discriminated durable identity for every session-backed subagent. Both modes carry the provider name. A `one-shot` descriptor optionally carries a caller-owned display `label`; a `continuable` descriptor requires the delegation `description` as its durable creation label and additionally snapshots resolved child `agentOptions.provider`/`model` and optional `persona`/`toolFilter` for cold resume. It never snapshots the merge-extensible `AgentOptions` object, so an unrelated extension value cannot break continuation and a later composition input is a deliberate version change. It omits `subagentDepth` (cold resume trusts the persisted header's `delegationDepth` as the monotone floor) and `outputSchema` (one run or Activation's result contract, not durable identity). +The descriptor (`SubagentDescriptorData` in [descriptor.ts](../../packages/subagent/subagent/src/descriptor.ts)) is a mode-discriminated durable identity for every session-backed subagent. Both modes carry the provider name. A `one-shot` descriptor optionally carries a caller-owned display `label`; a `continuable` descriptor requires the delegation `description` as its durable creation label and additionally snapshots resolved child `agentOptions.provider`/`model`/`reasoningEffort` and optional `persona`/`toolFilter` for cold resume. It never snapshots the merge-extensible `AgentOptions` object, so an unrelated extension value cannot break continuation and a later composition input is a deliberate version change. It omits `subagentDepth` (cold resume trusts the persisted header's `delegationDepth` as the monotone floor) and `outputSchema` (one run or Activation's result contract, not durable identity). A local one-shot provider appends the descriptor inside the child's initial turn before its first request. The continuation manager appends the descriptor after any provider-supplied lineage and before the initial prompt is admitted; `header.seedLength` remains the fork-lineage boundary: resume-time descriptor authority reads the child's own suffix, while the list-serving identity projection folds `subagent/descriptor` last-wins so the child's own descriptor overrides a fork-seeded ancestor's. The event is log-only: no `surfaceOp`, never in model history, and retained across compaction by the append-only log. Malformed current-version descriptors are corrupt; unsupported versions cannot be classified by this runtime. @@ -482,6 +489,22 @@ The spawn and fork backends create an ordinary one-shot agent through `parent.ct Generated from source by `scripts/gen-cordis-catalog.ts` (verified fresh by `pnpm run verify-cordis-catalog` in doc-sync; regenerate with `pnpm run gen-cordis-catalog`) — the language sides differ only in locale-specific paired document paths. Signature blocks use a `ts cordis-catalog` fence and keep the original source JSDoc; dispatch modes are defined in the [primer](../cordis-primer.md#dispatch-modes), and the framework-inherited `ctx` API lives in [cordis-api/inherited.md](../cordis-api/inherited.md). + + +### `ctx.subagentModelSelection` — `SubagentModelSelectionConfig` + +Singleton settings owner read by delegation tools when an Agent is published. + +```ts cordis-catalog +/** + * Read the preference for the next eligible Agent publication. + * @returns whether that Agent should receive model-selectable delegation. + */ +currentEnabled(): boolean +``` + +Source: [`packages/subagent/tool-subagent/src/model-selection-settings.ts`](../../packages/subagent/tool-subagent/src/model-selection-settings.ts) + ### `ctx.subagents` — `SubagentRuntime` diff --git a/docs/subsystems/subagent.zh.md b/docs/subsystems/subagent.zh.md index f397c19a0d..21b0dfd21d 100644 --- a/docs/subsystems/subagent.zh.md +++ b/docs/subsystems/subagent.zh.md @@ -25,6 +25,7 @@ Service Definition：[dsh-subagent](../../packages/subagent/subagent)（`ctx.sub * to `maxDepth`; the other names match. */ interface SubagentCapabilities { + readonly agentOptions: boolean readonly outputSchema: boolean readonly depthLimit: boolean readonly toolFilter: boolean @@ -34,7 +35,7 @@ interface SubagentCapabilities { ## 单次启动请求 -工具层根据模型输入和自身配置构建此请求；服务在 `start` 之前针对指定提供方进行校验。必填的 `parent` 提供会话 cwd、谱系与委派深度。可选的 output schema、depth、工具过滤器和 persona 需要对应的能力 flag 匹配。不支持的 schema 在启动时即失败；进程内后端将 filter 和 persona 的作用域限定在子 agent 创建阶段，并通过强制 capture 工具实现所支持的 object-rooted schema。 +工具层根据模型输入和自身配置构建此请求；服务在 `start` 之前针对指定提供方进行校验。必填的 `parent` 提供会话 cwd、谱系与委派深度。可选的 Agent 提供方、模型、推理强度与 token 覆盖、output schema、depth、工具过滤器和 persona 需要对应的能力 flag 匹配。进程内后端会把 `agentOptions` 合并到父 Agent 选项之上，将 filter 和 persona 的作用域限定在子 agent 创建阶段，并通过强制 capture 工具实现所支持的 object-rooted schema。当前进程外提供方会在启动其传输前拒绝 `agentOptions`。 ```ts type-equiv /** @@ -63,6 +64,12 @@ interface SubagentStartRequest { * remaining turn work when it fires afterward. */ readonly signal: AbortSignal + /** + * Optional host-Agent provider, model, reasoning-effort, and output-token + * overrides. Requires {@link SubagentCapabilities.agentOptions}; in-process + * providers merge them over the parent Agent's options when they create the + * child. + */ readonly agentOptions?: AgentOptions /** * Object-rooted JSON Schema within `assertObjectJsonSchema`'s enforced subset. Start rejects @@ -280,7 +287,7 @@ interface ContinuableCreateSpec { } ``` -描述符（[descriptor.ts](../../packages/subagent/subagent/src/descriptor.ts) 中的 `SubagentDescriptorData`）是每个由会话支撑的 subagent 所使用、按模式判别的持久化身份。两种模式都携带提供方名称。`one-shot` 描述符可以携带调用方拥有的可选显示 `label`；`continuable` 描述符要求以委派 `description` 作为持久化创建标签，并另外对已解析的子 agent `agentOptions.provider`／`model` 与可选的 `persona`／`toolFilter` 建立快照，用于冷恢复。它绝不会对可合并扩展的 `AgentOptions` 对象建立快照，因此无关的扩展值不会破坏继续执行，后续新增组合配置输入则是一次有意的版本更改。描述符省略 `subagentDepth`（冷恢复以持久化 header 中的 `delegationDepth` 作为单调下界）和 `outputSchema`（单次运行或 Activation 的结果约定，而非持久化身份）。 +描述符（[descriptor.ts](../../packages/subagent/subagent/src/descriptor.ts) 中的 `SubagentDescriptorData`）是每个由会话支撑的 subagent 所使用、按模式判别的持久化身份。两种模式都携带提供方名称。`one-shot` 描述符可以携带调用方拥有的可选显示 `label`；`continuable` 描述符要求以委派 `description` 作为持久化创建标签，并另外对已解析的子 agent `agentOptions.provider`／`model`／`reasoningEffort` 与可选的 `persona`／`toolFilter` 建立快照，用于冷恢复。它绝不会对可合并扩展的 `AgentOptions` 对象建立快照，因此无关的扩展值不会破坏继续执行，后续新增组合配置输入则是一次有意的版本更改。描述符省略 `subagentDepth`（冷恢复以持久化 header 中的 `delegationDepth` 作为单调下界）和 `outputSchema`（单次运行或 Activation 的结果约定，而非持久化身份）。 本地一次性提供方会在子 agent 的初始轮次内、首次请求前追加描述符。继续执行管理器会在任何提供方提供的谱系之后、初始提示词获准之前追加描述符；`header.seedLength` 仍是 fork 谱系边界：恢复时的描述符权威读取子 agent 自身的后缀，而供列表使用的身份投影以 last-wins 折叠 `subagent/descriptor`，子 agent 自己的描述符会覆盖 fork seed 中祖先的描述符。该事件只进入日志：不含 `surfaceOp`，绝不进入模型历史，并由仅追加日志跨压缩保留。格式错误的当前版本描述符属于损坏；本运行时无法对不受支持的版本进行分类。 @@ -486,6 +493,22 @@ spawn 和 fork 后端通过 `parent.ctx` 创建一个普通的单次 agent，将 Generated from source by `scripts/gen-cordis-catalog.ts` (verified fresh by `pnpm run verify-cordis-catalog` in doc-sync; regenerate with `pnpm run gen-cordis-catalog`) — the language sides differ only in locale-specific paired document paths. Signature blocks use a `ts cordis-catalog` fence and keep the original source JSDoc; dispatch modes are defined in the [primer](../cordis-primer.zh.md#dispatch-modes), and the framework-inherited `ctx` API lives in [cordis-api/inherited.md](../cordis-api/inherited.md). + + +### `ctx.subagentModelSelection` — `SubagentModelSelectionConfig` + +Singleton settings owner read by delegation tools when an Agent is published. + +```ts cordis-catalog +/** + * Read the preference for the next eligible Agent publication. + * @returns whether that Agent should receive model-selectable delegation. + */ +currentEnabled(): boolean +``` + +Source: [`packages/subagent/tool-subagent/src/model-selection-settings.ts`](../../packages/subagent/tool-subagent/src/model-selection-settings.ts) + ### `ctx.subagents` — `SubagentRuntime` diff --git a/docs/tool-catalog.i18n.yaml b/docs/tool-catalog.i18n.yaml index 10447d6107..38ad384da8 100644 --- a/docs/tool-catalog.i18n.yaml +++ b/docs/tool-catalog.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write docs/tool-catalog.md -tool-catalog.md: 1fa650f1e4e025274d069f27a6522abff46af2e2 -tool-catalog.zh.md: c3209e7007e9cf05770ccee0698f9e98a32e8363 +tool-catalog.md: 0cd8560a6851f0195e2272d1bd3b0bec2c171ac4 +tool-catalog.zh.md: cb225bc11afa6a6b022f2c7c104d4e1286f89260 diff --git a/docs/tool-catalog.md b/docs/tool-catalog.md index 1fa650f1e4..0cd8560a68 100644 --- a/docs/tool-catalog.md +++ b/docs/tool-catalog.md @@ -33,7 +33,7 @@ This table connects model-visible tool names to the plugin package and service s | `@deepseek-ai/dsh-tool-ralph` | `ralph` | `ctx.tools`, `ctx.workflowEngine`, `ctx.subagents`, `ctx.systemPrompt`, `a calling Agent (exec.agent parents every fresh round)` | `tool/call`, `tool/result`, `workflow and child session events during execution` | - | A fixed foreground workflow starts one fresh structured child per round; the model selects only the immutable objective and an optional round cap. | | `@deepseek-ai/dsh-tool-skill` | `skill` | `ctx.tools`, `ctx.agents`, `ctx.skills` | `tool/call`, `tool/result`, `user/message replacement catalogs via agent.inject()` | - | - | | `@deepseek-ai/dsh-tool-session-query` | `session_event_read`, `session_event_search`, `session_event_trace`, `session_search`, `session_trace` | `ctx.tools`, `ctx.systemPrompt`, `ctx.sessionQuery`, `a calling Agent for workspace authority` | `tool/call`, `tool/result` | - | The five read-only tools hide provider cursors and authorize every result from the immutable calling agent session. The package is opt-in; compositions that need enforced deadlines or bounded inline output also mount the generic timeout or spill policies. | -| `@deepseek-ai/dsh-tool-subagent` | `subagent` | `ctx.tools`, `ctx.subagents`, `ctx.systemPrompt` | `tool/call`, `tool/result`, `child session events through the chosen provider` | `subagent`, `subagent_fork` | The registered tool name is the load-time `toolName` config (default `subagent`); the schema above is that default. The shipped compositions load this package once per subagent backend, so the model additionally sees `subagent_fork` bound to the fork backend. Each instance's description, `run_in_background` parameter, and system-prompt policy follow its own `backgroundMode` and `enableRunInBackground`, so the two shipped schemas are not identical: `subagent` is `continuable` and defaults omitted calls to background with automatic settlement delivery, while `subagent_fork` stays `one-shot` and defaults them to foreground — see `packages/bundle/base/cordis.patch.yml` and `examples/acp-agent/cordis.yml`. | +| `@deepseek-ai/dsh-tool-subagent` | `list_subagent_models`, `subagent` | `ctx.tools`, `ctx.subagents`, `ctx.systemPrompt`, `ctx.llm for model discovery and selected-route validation` | `tool/call`, `tool/result`, `child session events through the chosen provider` | `subagent`, `subagent_fork` | The registered delegation name is the load-time `toolName` config (default `subagent`); the schema above shows static model selection enabled for reference. Model selection defaults off. Web presets sample the default-off Models preference for each new top-level Session and preserve that decision for its child Sessions; `subagent_fork` remains fixed-route. Explicit compositions may instead use static `enableModelSelection`. Each instance independently controls model selection, discovery ownership, and background behavior through `enableModelSelection`, `modelSelectionSettings`, `backgroundMode`, and `enableRunInBackground`. | | `@deepseek-ai/dsh-tool-subagent-control` | `interrupt_agent`, `list_agents`, `send_message` | `ctx.tools`, `ctx.subagents`, `ctx.agents and ctx.sessionProjections (list_agents only)` | `tool/call`, `tool/result`, `child session events through ctx.subagents` | - | The globally named control tools over continuable background subagents: provider-bound `tool-subagent` instances register distinct delegation tools, while this package registers `send_message` and `interrupt_agent` once, plus `list_agents` from its separately loaded `/list-agents` plugin (whose catalog rows use the sessionProjections and live Agent registries). | | `@deepseek-ai/dsh-tool-subagent-report` | `report` | `ctx.subagents`, `ctx.systemPrompt`, `a live continuable in-process child Agent` | `tool/call`, `tool/result`, `a user-role message in the direct parent session` | - | Registered per continuable in-process child rather than globally, so this schema is visible only inside such a child and survives its global `toolFilter`. The same contribution installs the child-scoped `tool:report` prompt section, which this catalog does not render. The parent-facing `send_message` tool is installed independently. | | `@deepseek-ai/dsh-tool-jobs` | `job_kill`, `job_list`, `job_output` | `ctx.tools`, `ctx.jobs`, `ctx.systemPrompt` | `tool/call`, `tool/result`, `user/message via agent.inject() for background completion notices` | - | The kind-agnostic background-job controller: background bash commands, PTY sends, and subagents are read, listed, and killed through the same three tools. Loading the plugin attaches the controller that arms producers' `ctx.jobs.start()`. | @@ -1501,9 +1501,31 @@ The five read-only tools hide provider cursors and authorize every result from t ## `@deepseek-ai/dsh-tool-subagent` +### `list_subagent_models` + +Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. + +```json +{ + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } +} +``` + +Source: [`packages/subagent/tool-subagent/src/list-models.ts`](../packages/subagent/tool-subagent/src/list-models.ts) + ### `subagent` -Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This call waits for the result by default. Set `run_in_background: true` to return a job id; collect with `job_output` and stop with `job_kill`. +Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This call waits for the result by default. Set `run_in_background: true` to return a job id; collect with `job_output` and stop with `job_kill`. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. ```json { @@ -1517,6 +1539,18 @@ Delegate a self-contained task to a subagent (a separate agent that works in its "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run as a background job and return its id. Defaults to false; collect with job_output or stop with job_kill." @@ -1531,7 +1565,7 @@ Delegate a self-contained task to a subagent (a separate agent that works in its Source: [`packages/subagent/tool-subagent/src/index.ts`](../packages/subagent/tool-subagent/src/index.ts) -The registered tool name is the load-time `toolName` config (default `subagent`); the schema above is that default. The shipped compositions load this package once per subagent backend, so the model additionally sees `subagent_fork` bound to the fork backend. Each instance's description, `run_in_background` parameter, and system-prompt policy follow its own `backgroundMode` and `enableRunInBackground`, so the two shipped schemas are not identical: `subagent` is `continuable` and defaults omitted calls to background with automatic settlement delivery, while `subagent_fork` stays `one-shot` and defaults them to foreground — see `packages/bundle/base/cordis.patch.yml` and `examples/acp-agent/cordis.yml`. +The registered delegation name is the load-time `toolName` config (default `subagent`); the schema above shows static model selection enabled for reference. Model selection defaults off. Web presets sample the default-off Models preference for each new top-level Session and preserve that decision for its child Sessions; `subagent_fork` remains fixed-route. Explicit compositions may instead use static `enableModelSelection`. Each instance independently controls model selection, discovery ownership, and background behavior through `enableModelSelection`, `modelSelectionSettings`, `backgroundMode`, and `enableRunInBackground`. diff --git a/docs/tool-catalog.zh.md b/docs/tool-catalog.zh.md index c3209e7007..cb225bc11a 100644 --- a/docs/tool-catalog.zh.md +++ b/docs/tool-catalog.zh.md @@ -37,7 +37,7 @@ | `@deepseek-ai/dsh-tool-ralph` | `ralph` | `ctx.tools`、`ctx.workflowEngine`、`ctx.subagents`、`ctx.systemPrompt`、`a calling Agent (exec.agent parents every fresh round)` | `tool/call`、`tool/result`、`workflow and child session events during execution` | - | 固定的前台工作流会在每个 Round 启动一个全新的结构化子级；模型只能选择不可变目标和可选的 Round 上限。 | | `@deepseek-ai/dsh-tool-skill` | `skill` | `ctx.tools`、`ctx.agents`、`ctx.skills` | `tool/call`、`tool/result`、`user/message replacement catalogs via agent.inject()` | - | - | | `@deepseek-ai/dsh-tool-session-query` | `session_event_read`、`session_event_search`、`session_event_trace`、`session_search`、`session_trace` | `ctx.tools`、`ctx.systemPrompt`、`ctx.sessionQuery`、`a calling Agent for workspace authority` | `tool/call`、`tool/result` | - | 这 5 个只读工具会隐藏提供方游标，并根据不可变的调用 agent 会话为每个结果授权。该包需要选择启用；需要强制截止时间或限制行内输出的组合还会挂载通用超时或 spill 策略。 | -| `@deepseek-ai/dsh-tool-subagent` | `subagent` | `ctx.tools`、`ctx.subagents`、`ctx.systemPrompt` | `tool/call`、`tool/result`、`child session events through the chosen provider` | `subagent`、`subagent_fork` | 注册的工具名称取决于加载时 `toolName` 配置（默认为 `subagent`）；上述 schema 对应默认值。随产品发布的组合会为每个 subagent 后端加载一次该包，因此模型还会看到绑定到 fork 后端的 `subagent_fork`。每个实例的描述、`run_in_background` 参数与 system prompt 策略取决于它自己的 `backgroundMode` 和 `enableRunInBackground`，因此两个随附 schema 并不相同：`subagent` 为 `continuable`，省略参数时默认后台运行，并由 runtime 自动投递结束结果；`subagent_fork` 保持 `one-shot`，省略参数时默认前台运行。详见 `packages/bundle/base/cordis.patch.yml` 和 `examples/acp-agent/cordis.yml`。 | +| `@deepseek-ai/dsh-tool-subagent` | `list_subagent_models`、`subagent` | `ctx.tools`、`ctx.subagents`、`ctx.systemPrompt`、`用于模型发现和所选路由校验的 ctx.llm` | `tool/call`、`tool/result`、`child session events through the chosen provider` | `subagent`、`subagent_fork` | 注册的委派工具名称取决于加载时 `toolName` 配置（默认为 `subagent`）；上述 schema 以静态启用模型选择作为参考。模型选择默认为关闭。Web preset 会在每个新顶层 Session 创建时读取 Models 页中默认关闭的偏好，并为其子 Session 保留该决定；`subagent_fork` 始终使用固定路由。显式组合也可以改用静态 `enableModelSelection`。每个实例通过 `enableModelSelection`、`modelSelectionSettings`、`backgroundMode` 与 `enableRunInBackground` 独立控制模型选择、发现工具持有权和后台行为。 | | `@deepseek-ai/dsh-tool-subagent-control` | `interrupt_agent`、`list_agents`、`send_message` | `ctx.tools`、`ctx.subagents`、`ctx.agents and ctx.sessionProjections (list_agents only)` | `tool/call`、`tool/result`、`child session events through ctx.subagents` | - | 这些是控制可继续后台 subagent 的全局命名工具：绑定提供方的 `tool-subagent` 实例注册不同的委派工具；本包注册一次 `send_message` 和 `interrupt_agent`，另由 `list_agents` 通过单独加载的 `/list-agents` 插件提供，其目录行使用 sessionProjections 和实时 Agent 注册表。 | | `@deepseek-ai/dsh-tool-subagent-report` | `report` | `ctx.subagents`、`ctx.systemPrompt`、`a live continuable in-process child Agent` | `tool/call`、`tool/result`、`a user-role message in the direct parent session` | - | 按可继续的进程内子级注册，而非全局注册，因此该 schema 仅在这种子级内部可见，并且不受其全局 `toolFilter` 影响。同一份贡献还会安装子级作用域的 `tool:report` 系统提示词 section，本目录不渲染该 section。面向父级的 `send_message` 工具单独安装。 | | `@deepseek-ai/dsh-tool-jobs` | `job_kill`、`job_list`、`job_output` | `ctx.tools`、`ctx.jobs`、`ctx.systemPrompt` | `tool/call`、`tool/result`、`user/message via agent.inject() for background completion notices` | - | 与任务种类无关的后台任务控制器：后台 bash 命令、PTY 发送和 subagent 都通过相同的 3 个工具读取、列出和终止。加载该插件会挂接控制器，从而启用生产方的 `ctx.jobs.start()`。 | @@ -1507,9 +1507,31 @@ lsp 工具将提供方选择和语言服务器子进程置于 ctx.lsp 之后， ## `@deepseek-ai/dsh-tool-subagent` +### `list_subagent_models` + +发现 subagent 可用的 LLM 路由，不更改当前 Agent。无参数调用会列出已注册提供方；提供 `provider` 时会列出其公布的模型；同时提供 `provider` 和 `model` 时会检查该精确模型及其推理强度。目录条目只提供建议：adapter 可能接受未列出的模型 id。把返回的 id 用于委派工具的 `provider`、`model` 与 `reasoning_effort` 字段。 + +```json +{ + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } +} +``` + +来源：[`packages/subagent/tool-subagent/src/list-models.ts`](../packages/subagent/tool-subagent/src/list-models.ts) + ### `subagent` -将一项自包含任务委派给 subagent（在自身上下文中工作的独立 agent），用它卸载聚焦且独立的工作，例如研究、限定范围的实现或分析，以免消耗当前对话的上下文。subagent 会返回结果，但不会返回中间步骤。请提供完整、独立的提示词，因为它看不到当前对话。此调用默认等待结果。设置 `run_in_background: true` 可返回 job id；使用 `job_output` 收集结果，使用 `job_kill` 停止任务。 +将一项自包含任务委派给 subagent（在自身上下文中工作的独立 agent），用它卸载聚焦且独立的工作，例如研究、限定范围的实现或分析，以免消耗当前对话的上下文。subagent 会返回结果，但不会返回中间步骤。请提供完整、独立的提示词，因为它看不到当前对话。此调用默认等待结果。设置 `run_in_background: true` 可返回 job id；使用 `job_output` 收集结果，使用 `job_kill` 停止任务。子级 LLM 选择是可选的。省略 `provider`、`model` 与 `reasoning_effort` 会使用配置的子级默认值，并从父 Agent 继承兼容的缺失值。先用 `list_subagent_models` 检查公布的路由和强度，再一起提供 `provider` 与 `model`。改变生效路由但不指定强度时，会使用所选模型的默认强度。 ```json { @@ -1523,6 +1545,18 @@ lsp 工具将提供方选择和语言服务器子进程置于 ctx.lsp 之后， "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run as a background job and return its id. Defaults to false; collect with job_output or stop with job_kill." @@ -1537,7 +1571,7 @@ lsp 工具将提供方选择和语言服务器子进程置于 ctx.lsp 之后， 来源：[`packages/subagent/tool-subagent/src/index.ts`](../packages/subagent/tool-subagent/src/index.ts) -注册的工具名称取决于加载时 `toolName` 配置（默认为 `subagent`）；上述 schema 对应默认值。随产品发布的组合会为每个 subagent 后端加载一次该包，因此模型还会看到绑定到 fork 后端的 `subagent_fork`。每个实例的描述、`run_in_background` 参数与 system prompt 策略取决于它自己的 `backgroundMode` 和 `enableRunInBackground`，因此两个随附 schema 并不相同：`subagent` 为 `continuable`，省略参数时默认后台运行，并由 runtime 自动投递结束结果；`subagent_fork` 保持 `one-shot`，省略参数时默认前台运行。详见 `packages/bundle/base/cordis.patch.yml` 和 `examples/acp-agent/cordis.yml`。 +注册的委派工具名称取决于加载时 `toolName` 配置（默认为 `subagent`）；上述 schema 以静态启用模型选择作为参考。模型选择默认为关闭。Web preset 会在每个新顶层 Session 创建时读取 Models 页中默认关闭的偏好，并为其子 Session 保留该决定；`subagent_fork` 始终使用固定路由。显式组合也可以改用静态 `enableModelSelection`。每个实例通过 `enableModelSelection`、`modelSelectionSettings`、`backgroundMode` 与 `enableRunInBackground` 独立控制模型选择、发现工具持有权和后台行为。 diff --git a/examples/acp-agent/cordis.yml b/examples/acp-agent/cordis.yml index ddfb914132..8f1df1955d 100644 --- a/examples/acp-agent/cordis.yml +++ b/examples/acp-agent/cordis.yml @@ -54,9 +54,17 @@ config: provider: spawn toolName: subagent + enableModelSelection: true backgroundMode: continuable maxDepth: 1 +# Fork omits model selection so provider/model stay equal to the parent and the +# inherited history remains eligible for KV Cache reuse. It stays one-shot because +# a continuable child's `report` tool and prompt section precede that history and +# invalidate the same prefix. `run_in_background` is off as an explicit foreground-only +# choice even though the shipped ACP profile mounts the generic Job runtime. +# See .agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md +# and .agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md. - id: tool-subagent-fork name: '@deepseek-ai/dsh-tool-subagent' config: diff --git a/examples/acp-agent/depth-two.cordis.snapshot.yml b/examples/acp-agent/depth-two.cordis.snapshot.yml index 89ef8c1035..245253690f 100644 --- a/examples/acp-agent/depth-two.cordis.snapshot.yml +++ b/examples/acp-agent/depth-two.cordis.snapshot.yml @@ -20,6 +20,7 @@ config: provider: spawn toolName: subagent + enableModelSelection: true backgroundMode: continuable maxDepth: 2 # Re-pin the recorded flash model for this scenario's corpus. diff --git a/examples/acp-agent/depth-two.cordis.yml b/examples/acp-agent/depth-two.cordis.yml index b2d4dbe039..d25b48fb60 100644 --- a/examples/acp-agent/depth-two.cordis.yml +++ b/examples/acp-agent/depth-two.cordis.yml @@ -5,5 +5,6 @@ config: provider: spawn toolName: subagent + enableModelSelection: true backgroundMode: continuable maxDepth: 2 diff --git a/examples/acp-agent/subagent-configured-effort.cordis.snapshot.yml b/examples/acp-agent/subagent-configured-effort.cordis.snapshot.yml new file mode 100644 index 0000000000..c89a77d49d --- /dev/null +++ b/examples/acp-agent/subagent-configured-effort.cordis.snapshot.yml @@ -0,0 +1,65 @@ +# Keyless counterpart to subagent-configured-effort.cordis.yml: disable the +# live adapter, insert replay, and apply the configured-effort rejection patch. +- id: llm-deepseek + name: '@deepseek-ai/dsh-llm-deepseek' + disabled: true + +- id: sandbox + name: '@deepseek-ai/dsh-sandbox-local' + config: + runnerCommand: + - bash + - -c + - while [ "$1" != "--" ]; do shift; done; shift; exec "$@" + - passthrough-runner + runnerFailureSignatures: + - 'passthrough-runner: profile rejected' + +- id: tool-subagent + name: '@deepseek-ai/dsh-tool-subagent' + config: + provider: spawn + toolName: subagent + enableModelSelection: true + backgroundMode: continuable + maxDepth: 1 + agentOptions: + provider: deepseek-official + model: deepseek-v4-flash + reasoningEffort: unsupported + +- id: acp + name: '@deepseek-ai/dsh-acp' + config: + provider: deepseek-official + model: deepseek-v4-flash + +- id: session-persistence-jsonl + name: '@deepseek-ai/dsh-session-persistence-jsonl' + config: + root: !!js process.env.DSH_SNAPSHOT_SESSIONS_ROOT ?? './.sessions' + compression: none + +- id: agent-instructions + name: '@deepseek-ai/dsh-agent-instructions' + config: + maxBytes: 65536 + +- id: system-prompt + name: '@deepseek-ai/dsh-system-prompt' + config: + persona: | + You are a coding assistant powered by the {{model}} model. Your working directory is {{cwd}}. Your bash tool runs under a file sandbox — a `[sandbox: file access denied …]` result is policy, not a command bug. + + Verify your work by running the code or tests. Keep answers brief and factual. + +- insert: + - id: llm-replay + name: '@deepseek-ai/dsh-llm-replay' + config: + providers: + - id: deepseek-official + name: DeepSeek + models: + - id: deepseek-v4-flash + - id: deepseek-v4-pro diff --git a/examples/acp-agent/subagent-configured-effort.cordis.yml b/examples/acp-agent/subagent-configured-effort.cordis.yml new file mode 100644 index 0000000000..9042fd3d91 --- /dev/null +++ b/examples/acp-agent/subagent-configured-effort.cordis.yml @@ -0,0 +1,15 @@ +# Configured-effort rejection snapshot overlay: keep one invalid configured +# effort so the tool rejects before starting a child instead of deferring the +# failure to the child agent loop. +- id: tool-subagent + name: '@deepseek-ai/dsh-tool-subagent' + config: + provider: spawn + toolName: subagent + enableModelSelection: true + backgroundMode: continuable + maxDepth: 1 + agentOptions: + provider: deepseek-official + model: deepseek-v4-flash + reasoningEffort: unsupported diff --git a/examples/acp-agent/tests/acp.snapshot.ts b/examples/acp-agent/tests/acp.snapshot.ts index f6bd6124e7..73286b97a6 100644 --- a/examples/acp-agent/tests/acp.snapshot.ts +++ b/examples/acp-agent/tests/acp.snapshot.ts @@ -70,6 +70,9 @@ const SUBAGENT_DURABILITY_FAILURE_CONFIG = fileURLToPath( const SUBAGENT_CONTINUABLE_INHERITANCE_CONFIG = fileURLToPath( new URL('../subagent-continuable-inheritance.cordis.yml', import.meta.url), ) +const SUBAGENT_CONFIGURED_EFFORT_CONFIG = fileURLToPath( + new URL('../subagent-configured-effort.cordis.yml', import.meta.url), +) const LSP_CONFIG = fileURLToPath(new URL('./lsp.cordis.yml', import.meta.url)) const WEB_CONFIG = fileURLToPath(new URL('../web.cordis.yml', import.meta.url)) const FS_SEARCH_CONFIG = fileURLToPath(new URL('./fs-search.cordis.yml', import.meta.url)) @@ -201,6 +204,10 @@ const SCENARIOS: Scenario[] = [ hasModelTurn: true, recorded: false, overridden: true, + pinsHeader: true, + headerClass: 'session-title', + systemPromptSource: 'text-turn', + toolSchemasSource: 'text-turn', configPath: SESSION_TITLE_CONFIG, }, { name: 'tool-call-turn', hasModelTurn: true, recorded: true }, @@ -402,8 +409,8 @@ const SCENARIOS: Scenario[] = [ // An overwrite whose replacement is at/above the configured diff-basis bound: // the persisted result meta carries no contextual hunks and presentation // falls back to the whole-file diff. The overlay leaves the prompt and tool - // sequence identical to text-turn, but the freshly recorded header carries - // the current adapter capability fields, so the scenario pins its own class. + // sequence identical to text-turn. The scenario pins its own header class + // for config fields while sharing the unchanged prompt and tool schema. { name: 'fs-write-overwrite-bounded', hasModelTurn: true, @@ -559,6 +566,20 @@ const SCENARIOS: Scenario[] = [ overridden: true, configPath: DEPTH_TWO_CONFIG, }, + // Authored keyless replay first discovers the exact child route through the + // assembled directory tool, then proves a configured effort is validated + // before child creation through the same live replay LLM registry. + { + name: 'subagent-configured-effort-rejection', + hasModelTurn: true, + recorded: false, + overridden: true, + pinsHeader: true, + headerClass: 'subagent-configured-effort', + systemPromptSource: 'text-turn', + toolSchemasSource: 'text-turn', + configPath: SUBAGENT_CONFIGURED_EFFORT_CONFIG, + }, // Authored keyless replay through the assembled app: a one-shot child calls // the real ask_user_question tool, the runtime-ownership guard rejects before // the tripwire provider, and the child carries the unresolved decision in its diff --git a/examples/acp-agent/tests/snapshots/advanced-toolchain/session.1.jsonl b/examples/acp-agent/tests/snapshots/advanced-toolchain/session.1.jsonl index 01ba355631..607c9b3bb3 100644 --- a/examples/acp-agent/tests/snapshots/advanced-toolchain/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/advanced-toolchain/session.1.jsonl @@ -5,10 +5,10 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly DIRECT_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"ebe0cfa0-a909-47e0-8294-28ad84a8fe77"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Check direct child"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Check direct child"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly DIRECT_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"ebe0cfa0-a909-47e0-8294-28ad84a8fe77"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"e3c23441-606f-4e7a-8338-b434c0d04a4e"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"b1814e62-f9de-49fc-8e60-4271eecb3500"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly DIRECT_CHILD_OK and","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} diff --git a/examples/acp-agent/tests/snapshots/advanced-toolchain/session.2.jsonl b/examples/acp-agent/tests/snapshots/advanced-toolchain/session.2.jsonl index 73958c517c..d3cbf0e856 100644 --- a/examples/acp-agent/tests/snapshots/advanced-toolchain/session.2.jsonl +++ b/examples/acp-agent/tests/snapshots/advanced-toolchain/session.2.jsonl @@ -5,10 +5,10 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly WORKFLOW_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"2ac2cc54-9bce-4cfa-a569-a64f51bc30a7"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly WORKFLOW_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"2ac2cc54-9bce-4cfa-a569-a64f51bc30a7"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"5d128e81-c7c2-4cd0-ad1c-7409b33650fc"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"f82215a6-9c52-4c75-b46b-f722a1b64f72"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly WORKFLOW_CHILD_OK and","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} diff --git a/examples/acp-agent/tests/snapshots/advanced-toolchain/system-prompt.expected.md b/examples/acp-agent/tests/snapshots/advanced-toolchain/system-prompt.expected.md index 1743643d95..b880f67453 100644 --- a/examples/acp-agent/tests/snapshots/advanced-toolchain/system-prompt.expected.md +++ b/examples/acp-agent/tests/snapshots/advanced-toolchain/system-prompt.expected.md @@ -301,6 +301,13 @@ interface ToolArgsMap { /** children (default) lists direct children only; descendants walks the complete tree below you. */ scope?: "children" | "descendants"; } & Record; + /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */ + list_subagent_models: { + /** Registered LLM provider id. Omit to list providers. */ + provider?: string; + /** Exact model id to inspect. Requires provider; omit to list that provider's advertised models. */ + model?: string; + } & Record; /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */ ralph: { /** The immutable completion objective for every fresh Ralph round. */ @@ -351,12 +358,18 @@ interface ToolArgsMap { /** Optional parameter of `view` command when `path` points to a file. If none is given, the full file is shown. If provided, the file will be shown in the indicated line number range, e.g. [11, 12] will show lines 11 and 12. Indexing at 1 to start. Setting `[start_line, -1]` shows all lines from `start_line` to the end of the file. */ view_range?: number[]; } & Record; - /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. */ + /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */ subagent: { /** A short (3-5 word) description of the delegated task, for display. */ description: string; /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */ prompt: string; + /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */ + provider?: string; + /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */ + model?: string; + /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */ + reasoning_effort?: string; /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */ run_in_background?: boolean; } & Record; @@ -587,6 +600,7 @@ interface ToolOutputMap { parent?: string; depth?: number; })[]; + list_subagent_models: string; ralph: { runId: string; agentsStarted: number; diff --git a/examples/acp-agent/tests/snapshots/advanced-toolchain/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/advanced-toolchain/tool-schemas.expected.json index dcce863f94..2f1950a691 100644 --- a/examples/acp-agent/tests/snapshots/advanced-toolchain/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/advanced-toolchain/tool-schemas.expected.json @@ -457,6 +457,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -627,7 +644,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -639,6 +656,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/both-mode-turn/system-prompt.expected.md b/examples/acp-agent/tests/snapshots/both-mode-turn/system-prompt.expected.md index 37df6287ed..b667c8dd6b 100644 --- a/examples/acp-agent/tests/snapshots/both-mode-turn/system-prompt.expected.md +++ b/examples/acp-agent/tests/snapshots/both-mode-turn/system-prompt.expected.md @@ -134,6 +134,13 @@ interface ToolArgsMap { /** children (default) lists direct children only; descendants walks the complete tree below you. */ scope?: "children" | "descendants"; } & Record; + /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */ + list_subagent_models: { + /** Registered LLM provider id. Omit to list providers. */ + provider?: string; + /** Exact model id to inspect. Requires provider; omit to list that provider's advertised models. */ + model?: string; + } & Record; /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */ ralph: { /** The immutable completion objective for every fresh Ralph round. */ @@ -184,12 +191,18 @@ interface ToolArgsMap { /** Optional parameter of `view` command when `path` points to a file. If none is given, the full file is shown. If provided, the file will be shown in the indicated line number range, e.g. [11, 12] will show lines 11 and 12. Indexing at 1 to start. Setting `[start_line, -1]` shows all lines from `start_line` to the end of the file. */ view_range?: number[]; } & Record; - /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. */ + /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */ subagent: { /** A short (3-5 word) description of the delegated task, for display. */ description: string; /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */ prompt: string; + /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */ + provider?: string; + /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */ + model?: string; + /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */ + reasoning_effort?: string; /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */ run_in_background?: boolean; } & Record; @@ -401,6 +414,7 @@ interface ToolOutputMap { parent?: string; depth?: number; })[]; + list_subagent_models: string; ralph: { runId: string; agentsStarted: number; diff --git a/examples/acp-agent/tests/snapshots/both-mode-turn/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/both-mode-turn/tool-schemas.expected.json index b8a2ed790f..bf85198220 100644 --- a/examples/acp-agent/tests/snapshots/both-mode-turn/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/both-mode-turn/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -430,7 +447,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -442,6 +459,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/code-mode-read-image/system-prompt.expected.md b/examples/acp-agent/tests/snapshots/code-mode-read-image/system-prompt.expected.md index daf622df60..c9bad7d1fa 100644 --- a/examples/acp-agent/tests/snapshots/code-mode-read-image/system-prompt.expected.md +++ b/examples/acp-agent/tests/snapshots/code-mode-read-image/system-prompt.expected.md @@ -136,6 +136,13 @@ interface ToolArgsMap { /** children (default) lists direct children only; descendants walks the complete tree below you. */ scope?: "children" | "descendants"; } & Record; + /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */ + list_subagent_models: { + /** Registered LLM provider id. Omit to list providers. */ + provider?: string; + /** Exact model id to inspect. Requires provider; omit to list that provider's advertised models. */ + model?: string; + } & Record; /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */ ralph: { /** The immutable completion objective for every fresh Ralph round. */ @@ -186,12 +193,18 @@ interface ToolArgsMap { /** Optional parameter of `view` command when `path` points to a file. If none is given, the full file is shown. If provided, the file will be shown in the indicated line number range, e.g. [11, 12] will show lines 11 and 12. Indexing at 1 to start. Setting `[start_line, -1]` shows all lines from `start_line` to the end of the file. */ view_range?: number[]; } & Record; - /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. */ + /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */ subagent: { /** A short (3-5 word) description of the delegated task, for display. */ description: string; /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */ prompt: string; + /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */ + provider?: string; + /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */ + model?: string; + /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */ + reasoning_effort?: string; /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */ run_in_background?: boolean; } & Record; @@ -403,6 +416,7 @@ interface ToolOutputMap { parent?: string; depth?: number; })[]; + list_subagent_models: string; ralph: { runId: string; agentsStarted: number; diff --git a/examples/acp-agent/tests/snapshots/code-mode-turn/system-prompt.expected.md b/examples/acp-agent/tests/snapshots/code-mode-turn/system-prompt.expected.md index 6894f13fb6..7506ad8373 100644 --- a/examples/acp-agent/tests/snapshots/code-mode-turn/system-prompt.expected.md +++ b/examples/acp-agent/tests/snapshots/code-mode-turn/system-prompt.expected.md @@ -136,6 +136,13 @@ interface ToolArgsMap { /** children (default) lists direct children only; descendants walks the complete tree below you. */ scope?: "children" | "descendants"; } & Record; + /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */ + list_subagent_models: { + /** Registered LLM provider id. Omit to list providers. */ + provider?: string; + /** Exact model id to inspect. Requires provider; omit to list that provider's advertised models. */ + model?: string; + } & Record; /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */ ralph: { /** The immutable completion objective for every fresh Ralph round. */ @@ -186,12 +193,18 @@ interface ToolArgsMap { /** Optional parameter of `view` command when `path` points to a file. If none is given, the full file is shown. If provided, the file will be shown in the indicated line number range, e.g. [11, 12] will show lines 11 and 12. Indexing at 1 to start. Setting `[start_line, -1]` shows all lines from `start_line` to the end of the file. */ view_range?: number[]; } & Record; - /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. */ + /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */ subagent: { /** A short (3-5 word) description of the delegated task, for display. */ description: string; /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */ prompt: string; + /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */ + provider?: string; + /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */ + model?: string; + /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */ + reasoning_effort?: string; /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */ run_in_background?: boolean; } & Record; @@ -403,6 +416,7 @@ interface ToolOutputMap { parent?: string; depth?: number; })[]; + list_subagent_models: string; ralph: { runId: string; agentsStarted: number; diff --git a/examples/acp-agent/tests/snapshots/fs-glob-sampling/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/fs-glob-sampling/tool-schemas.expected.json index 993a7579bd..2819e54870 100644 --- a/examples/acp-agent/tests/snapshots/fs-glob-sampling/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/fs-glob-sampling/tool-schemas.expected.json @@ -180,6 +180,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -313,7 +330,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -325,6 +342,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/lsp-definition/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/lsp-definition/tool-schemas.expected.json index 4ea490a884..c012852f0a 100644 --- a/examples/acp-agent/tests/snapshots/lsp-definition/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/lsp-definition/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "lsp", "description": "Query a language server for precise code navigation. operation is one of goToDefinition, findReferences, goToImplementation, hover. line and character are one-based UTF-16 cursor coordinates. findReferences includes the declaration.", @@ -446,7 +463,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -458,6 +475,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/product-subagent-both/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/product-subagent-both/tool-schemas.expected.json index e1d954e615..5eec9bb706 100644 --- a/examples/acp-agent/tests/snapshots/product-subagent-both/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/product-subagent-both/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -409,7 +426,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -421,6 +438,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/product-subagent-codex/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/product-subagent-codex/tool-schemas.expected.json index 944a002e53..2d5b27c48b 100644 --- a/examples/acp-agent/tests/snapshots/product-subagent-codex/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/product-subagent-codex/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -409,7 +426,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -421,6 +438,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/product-subagent-result-diagnostic/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/product-subagent-result-diagnostic/tool-schemas.expected.json index 0ed6e087db..bb5b4b7411 100644 --- a/examples/acp-agent/tests/snapshots/product-subagent-result-diagnostic/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/product-subagent-result-diagnostic/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -409,7 +426,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -421,6 +438,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/pty-tools/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/pty-tools/tool-schemas.expected.json index 6ab4f41978..2303325732 100644 --- a/examples/acp-agent/tests/snapshots/pty-tools/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/pty-tools/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -409,7 +426,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -421,6 +438,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/session-query-spill/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/session-query-spill/tool-schemas.expected.json index c423cdb57c..7ff41194b3 100644 --- a/examples/acp-agent/tests/snapshots/session-query-spill/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/session-query-spill/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -613,7 +630,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -625,6 +642,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/subagent-child-question-rejection/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-child-question-rejection/session.1.jsonl index 323cdcdb4c..e6f032662c 100644 --- a/examples/acp-agent/tests/snapshots/subagent-child-question-rejection/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-child-question-rejection/session.1.jsonl @@ -5,7 +5,7 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Call ask_user_question once to ask whether deployment should use the CUDA fallback. If the tool returns an error, include the unresolved question verbatim in your final result."}],"source":{"kind":"user"},"role":"user","id":"106c2785-219e-46e8-8386-497ac6a98f68"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Check deployment question"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Check deployment question"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Call ask_user_question once to ask whether deployment should use the CUDA fallback. If the tool returns an error, include the unresolved question verbatim in your final result."}],"source":{"kind":"user"},"role":"user","id":"106c2785-219e-46e8-8386-497ac6a98f68"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"d8734c8a-d956-4e3f-8d28-399adf51a203"},"surfaceOp":"append"} diff --git a/examples/acp-agent/tests/snapshots/subagent-child-question-rejection/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/subagent-child-question-rejection/tool-schemas.expected.json index fdda53355f..6aa10aa7d5 100644 --- a/examples/acp-agent/tests/snapshots/subagent-child-question-rejection/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/subagent-child-question-rejection/tool-schemas.expected.json @@ -323,6 +323,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -472,7 +489,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -484,6 +501,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/input.json b/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/input.json new file mode 100644 index 0000000000..eb1c2ee7ff --- /dev/null +++ b/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/input.json @@ -0,0 +1,7 @@ +{ + "steps": [ + { "op": "initialize" }, + { "op": "newSession" }, + { "op": "prompt", "text": "Inspect the configured child model, then attempt one subagent call so its configured reasoning effort is validated before child creation." } + ] +} diff --git a/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/replay.override.json b/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/replay.override.json new file mode 100644 index 0000000000..0ac208bb71 --- /dev/null +++ b/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/replay.override.json @@ -0,0 +1,32 @@ +[ + { + "kind": "chunks", + "chunks": [ + { "type": "block-start", "index": 0, "blockType": "tool-call" }, + { "type": "tool-call-delta", "index": 0, "id": "call_list_child_model", "name": "list_subagent_models", "argumentsDelta": "{\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\"}" }, + { "type": "block-end", "index": 0, "block": { "type": "tool-call", "id": "call_list_child_model", "name": "list_subagent_models", "arguments": "{\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\"}" } }, + { "type": "usage", "usage": { "inputTokens": 10, "outputTokens": 5 } }, + { "type": "finish", "reason": { "kind": "tool-calls" } } + ] + }, + { + "kind": "chunks", + "chunks": [ + { "type": "block-start", "index": 0, "blockType": "tool-call" }, + { "type": "tool-call-delta", "index": 0, "id": "call_configured_effort", "name": "subagent", "argumentsDelta": "{\"description\":\"Validate configured effort\",\"prompt\":\"This child must never start.\",\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\",\"run_in_background\":false}" }, + { "type": "block-end", "index": 0, "block": { "type": "tool-call", "id": "call_configured_effort", "name": "subagent", "arguments": "{\"description\":\"Validate configured effort\",\"prompt\":\"This child must never start.\",\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\",\"run_in_background\":false}" } }, + { "type": "usage", "usage": { "inputTokens": 10, "outputTokens": 5 } }, + { "type": "finish", "reason": { "kind": "tool-calls" } } + ] + }, + { + "kind": "chunks", + "chunks": [ + { "type": "block-start", "index": 0, "blockType": "text" }, + { "type": "text-delta", "index": 0, "text": "CONFIGURED_EFFORT_REJECTED" }, + { "type": "block-end", "index": 0, "block": { "type": "text", "text": "CONFIGURED_EFFORT_REJECTED" } }, + { "type": "usage", "usage": { "inputTokens": 10, "outputTokens": 2 } }, + { "type": "finish", "reason": { "kind": "stop" } } + ] + } +] diff --git a/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/session.jsonl b/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/session.jsonl new file mode 100644 index 0000000000..25f746a2c1 --- /dev/null +++ b/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/session.jsonl @@ -0,0 +1,41 @@ +{"type":"session","version":0,"id":"44444444-4444-4444-8444-444444444444","createdAt":1000,"cwd":"{{cwd}}","delegationDepth":0} +{"type":"permission/preset","data":{"preset":"danger-full-access"}} +{"type":"sandbox/mode","data":{"mode":"danger-full-access"}} +{"type":"approval/policy","data":{"policy":"never"}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Inspect the configured child model, then attempt one subagent call so its configured reasoning effort is validated before child creation."}],"source":{"kind":"user"},"role":"user","id":"04d53124-d38b-4705-a711-d0b5755b74c9"}]}} +{"type":"turn/start","data":{"turn":1}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} +{"type":"step/start","data":{"turn":1,"step":1}} +{"type":"user/message","data":{"content":[{"type":"text","text":"Inspect the configured child model, then attempt one subagent call so its configured reasoning effort is validated before child creation."}],"source":{"kind":"user"},"role":"user","id":"04d53124-d38b-4705-a711-d0b5755b74c9"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."}]},"role":"user","id":"55365caf-6fcc-484b-a4b7-646914654bbc"},"surfaceOp":"append"} +{"type":"session/title","data":{"title":"Inspect the configured child model,","messageSeqs":[7],"source":{"kind":"fallback"}}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} +{"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} +{"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} +{"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"tool-call-delta","index":0,"id":"call_list_child_model","name":"list_subagent_models","argumentsDelta":"{\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\"}"}}} +{"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"call_list_child_model","name":"list_subagent_models","arguments":"{\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\"}"}}}} +{"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} +{"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} +{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"call_list_child_model","name":"list_subagent_models","arguments":"{\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"000adf36-9918-4ad3-ab38-2fdb9003d0d7"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[12,13,14,15,16],"surfaceOp":"append"} +{"type":"tool/call","data":{"turn":1,"step":1,"callId":"call_list_child_model","name":"list_subagent_models","arguments":"{\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\"}"}} +{"type":"tool/result","data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"call_list_child_model"},"content":[{"type":"tool-result","toolCallId":"call_list_child_model","content":[{"type":"text","text":"deepseek-official/deepseek-v4-flash — deepseek-v4-flash\nReasoning efforts:\n(no advertised reasoning efforts)"}],"isError":false}],"role":"user","id":"f480c34c-cd56-4b8e-be7c-e6ce50d3f70a"}},"sourceEventSeqs":[18],"surfaceOp":"append"} +{"type":"step/end","data":{"turn":1,"step":1}} +{"type":"step/start","data":{"turn":1,"step":2}} +{"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} +{"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"tool-call-delta","index":0,"id":"call_configured_effort","name":"subagent","argumentsDelta":"{\"description\":\"Validate configured effort\",\"prompt\":\"This child must never start.\",\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\",\"run_in_background\":false}"}}} +{"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"call_configured_effort","name":"subagent","arguments":"{\"description\":\"Validate configured effort\",\"prompt\":\"This child must never start.\",\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\",\"run_in_background\":false}"}}}} +{"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} +{"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} +{"type":"assistant/message","data":{"turn":1,"step":2,"message":{"role":"assistant","content":[{"type":"tool-call","id":"call_configured_effort","name":"subagent","arguments":"{\"description\":\"Validate configured effort\",\"prompt\":\"This child must never start.\",\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\",\"run_in_background\":false}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"aec67548-de79-480f-8aec-2bd458ddbb41"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[22,23,24,25,26],"surfaceOp":"append"} +{"type":"tool/call","data":{"turn":1,"step":2,"callId":"call_configured_effort","name":"subagent","arguments":"{\"description\":\"Validate configured effort\",\"prompt\":\"This child must never start.\",\"provider\":\"deepseek-official\",\"model\":\"deepseek-v4-flash\",\"run_in_background\":false}"}} +{"type":"tool/result","data":{"turn":1,"step":2,"message":{"source":{"kind":"tool","callId":"call_configured_effort"},"content":[{"type":"tool-result","toolCallId":"call_configured_effort","content":[{"type":"text","text":"Error: provider \"deepseek-official\" model \"deepseek-v4-flash\" does not support reasoning effort \"unsupported\""}],"isError":true}],"role":"user","id":"9a59cf14-e222-40de-85f4-6a4ab4f5a177"},"error":{"name":"LlmError","code":"UNSUPPORTED_REASONING_EFFORT"}},"sourceEventSeqs":[28],"surfaceOp":"append"} +{"type":"step/end","data":{"turn":1,"step":2}} +{"type":"step/start","data":{"turn":1,"step":3}} +{"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} +{"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"text-delta","index":0,"text":"CONFIGURED_EFFORT_REJECTED"}}} +{"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"CONFIGURED_EFFORT_REJECTED"}}}} +{"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":2}}}} +{"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} +{"type":"assistant/message","data":{"turn":1,"step":3,"message":{"role":"assistant","content":[{"type":"text","text":"CONFIGURED_EFFORT_REJECTED"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"5a1911c6-f487-458f-b802-4a66221ec047"},"usage":{"inputTokens":10,"outputTokens":2}},"sourceEventSeqs":[32,33,34,35,36],"surfaceOp":"append"} +{"type":"step/end","data":{"turn":1,"step":3}} +{"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} diff --git a/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/stdout.expected.jsonl b/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/stdout.expected.jsonl new file mode 100644 index 0000000000..7d5829fb9d --- /dev/null +++ b/examples/acp-agent/tests/snapshots/subagent-configured-effort-rejection/stdout.expected.jsonl @@ -0,0 +1,8 @@ +{"jsonrpc":"2.0","id":1,"result":{"protocolVersion":1,"agentInfo":{"name":"deepseek-harness-acp","version":"0.0.1"},"agentCapabilities":{"mcpCapabilities":{"http":true},"promptCapabilities":{"image":false,"audio":false,"embeddedContext":false},"sessionCapabilities":{"close":{},"list":{},"resume":{}}},"authMethods":[]}} +{"jsonrpc":"2.0","id":2,"result":{"sessionId":"{{sessionId}}","configOptions":[{"id":"model","name":"Model","category":"model","type":"select","currentValue":"[\"deepseek-official\",\"deepseek-v4-flash\"]","options":[{"group":"deepseek-official","name":"DeepSeek","options":[{"value":"[\"deepseek-official\",\"deepseek-v4-flash\"]","name":"deepseek-v4-flash"},{"value":"[\"deepseek-official\",\"deepseek-v4-pro\"]","name":"deepseek-v4-pro"}]}]}]}} +{"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"tool_call","toolCallId":"call_list_child_model","title":"list_subagent_models","kind":"other","status":"in_progress","rawInput":{"provider":"deepseek-official","model":"deepseek-v4-flash"}}}} +{"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"tool_call_update","toolCallId":"call_list_child_model","status":"completed","content":[{"type":"content","content":{"type":"text","text":"deepseek-official/deepseek-v4-flash — deepseek-v4-flash\nReasoning efforts:\n(no advertised reasoning efforts)"}}]}}} +{"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"tool_call","toolCallId":"call_configured_effort","title":"subagent","kind":"other","status":"in_progress","rawInput":{"description":"Validate configured effort","prompt":"This child must never start.","provider":"deepseek-official","model":"deepseek-v4-flash","run_in_background":false}}}} +{"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"tool_call_update","toolCallId":"call_configured_effort","status":"failed","content":[{"type":"content","content":{"type":"text","text":"Error: provider \"deepseek-official\" model \"deepseek-v4-flash\" does not support reasoning effort \"unsupported\""}}]}}} +{"jsonrpc":"2.0","method":"session/update","params":{"sessionId":"{{sessionId}}","update":{"sessionUpdate":"agent_message_chunk","messageId":"{{messageId}}","content":{"type":"text","text":"CONFIGURED_EFFORT_REJECTED"}}}} +{"jsonrpc":"2.0","id":3,"result":{"stopReason":"end_turn"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-continuable-inheritance/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-continuable-inheritance/session.1.jsonl index 841bdb291d..d2d2f47853 100644 --- a/examples/acp-agent/tests/snapshots/subagent-continuable-inheritance/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-continuable-inheritance/session.1.jsonl @@ -1,5 +1,5 @@ {"type":"session","version":0,"id":"33333333-3333-4333-8333-333333333333","createdAt":1789000001000,"cwd":"{{cwd}}","parentSession":"11111111-1111-4111-8111-111111111111","origin":"subagent","delegationDepth":1} -{"type":"subagent/descriptor","data":{"version":2,"mode":"continuable","provider":"spawn","label":"Reply with CHILD_OK","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"continuable","provider":"spawn","label":"Reply with CHILD_OK","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} {"type":"session/end-seed","data":{}} {"type":"sandbox/mode","data":{"mode":"read-only","source":"delegation"}} {"type":"approval/policy","data":{"policy":"never","source":"delegation"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-continuable-inheritance/tool-schemas.1.expected.json b/examples/acp-agent/tests/snapshots/subagent-continuable-inheritance/tool-schemas.1.expected.json index 38f4eae1ad..62937be9b1 100644 --- a/examples/acp-agent/tests/snapshots/subagent-continuable-inheritance/tool-schemas.1.expected.json +++ b/examples/acp-agent/tests/snapshots/subagent-continuable-inheritance/tool-schemas.1.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -425,7 +442,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -437,6 +454,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/subagent-continuable/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-continuable/session.1.jsonl index a03acfdecb..685168de6b 100644 --- a/examples/acp-agent/tests/snapshots/subagent-continuable/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-continuable/session.1.jsonl @@ -1,5 +1,5 @@ {"type":"session","version":0,"id":"33333333-3333-4333-8333-333333333333","createdAt":1789000001000,"cwd":"{{cwd}}","parentSession":"11111111-1111-4111-8111-111111111111","origin":"subagent","delegationDepth":1} -{"type":"subagent/descriptor","data":{"version":2,"mode":"continuable","provider":"spawn","label":"Reply with CHILD_OK","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"continuable","provider":"spawn","label":"Reply with CHILD_OK","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} {"type":"session/end-seed","data":{}} {"type":"sandbox/mode","data":{"mode":"danger-full-access","source":"delegation"}} {"type":"approval/policy","data":{"policy":"never","source":"delegation"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-continuable/tool-schemas.1.expected.json b/examples/acp-agent/tests/snapshots/subagent-continuable/tool-schemas.1.expected.json index 38f4eae1ad..62937be9b1 100644 --- a/examples/acp-agent/tests/snapshots/subagent-continuable/tool-schemas.1.expected.json +++ b/examples/acp-agent/tests/snapshots/subagent-continuable/tool-schemas.1.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -425,7 +442,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -437,6 +454,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/subagent-depth-two-rejection/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-depth-two-rejection/session.1.jsonl index b64494a028..d59385affa 100644 --- a/examples/acp-agent/tests/snapshots/subagent-depth-two-rejection/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-depth-two-rejection/session.1.jsonl @@ -5,10 +5,10 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Call subagent once. Ask that child to attempt one further subagent call, then report the result."}],"source":{"kind":"user"},"role":"user","id":"a8129357-1bde-4cbd-90b4-6b8ad51d52e1"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Start depth one"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Start depth one"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Call subagent once. Ask that child to attempt one further subagent call, then report the result."}],"source":{"kind":"user"},"role":"user","id":"a8129357-1bde-4cbd-90b4-6b8ad51d52e1"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"72272791-eefd-48f8-94da-02b132ae9d2a"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"f544ed7b-5a1f-4b6e-93b5-6af8342385fc"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Call subagent once. Ask that","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-depth-two-rejection/session.2.jsonl b/examples/acp-agent/tests/snapshots/subagent-depth-two-rejection/session.2.jsonl index 3e3dfa357b..362fed6c4e 100644 --- a/examples/acp-agent/tests/snapshots/subagent-depth-two-rejection/session.2.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-depth-two-rejection/session.2.jsonl @@ -5,10 +5,10 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Attempt one subagent call beyond the configured cap, then report the rejection."}],"source":{"kind":"user"},"role":"user","id":"d4dc5a16-e542-4dd9-8e82-e6b7829cfc4b"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Start depth two"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Start depth two"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Attempt one subagent call beyond the configured cap, then report the rejection."}],"source":{"kind":"user"},"role":"user","id":"d4dc5a16-e542-4dd9-8e82-e6b7829cfc4b"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"c54120cc-6a7f-41f6-a71d-42b4805fa2ca"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"5d344fef-f707-49ea-b804-ac384bf52700"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Attempt one subagent call beyond","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-fork-in-process/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-fork-in-process/session.1.jsonl index 57bbed4d87..36c2654033 100644 --- a/examples/acp-agent/tests/snapshots/subagent-fork-in-process/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-fork-in-process/session.1.jsonl @@ -28,7 +28,7 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"What is the project codeword mentioned earlier in this conversation? Reply with exactly that one word and nothing else."}],"source":{"kind":"user"},"role":"user","id":"d037163e-ed56-4c9c-b5d1-57df017d618c"}]}} {"type":"turn/start","data":{"turn":2}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"fork","label":"Recall project codeword"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"fork","label":"Recall project codeword"}} {"type":"step/start","data":{"turn":2,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"What is the project codeword mentioned earlier in this conversation? Reply with exactly that one word and nothing else."}],"source":{"kind":"user"},"role":"user","id":"d037163e-ed56-4c9c-b5d1-57df017d618c"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"257e572f-6f95-48f9-b3d7-4ea8b162f374"},"surfaceOp":"append"} diff --git a/examples/acp-agent/tests/snapshots/subagent-list-agents/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-list-agents/session.1.jsonl index b80c3d4936..2e6847126b 100644 --- a/examples/acp-agent/tests/snapshots/subagent-list-agents/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-list-agents/session.1.jsonl @@ -1,5 +1,5 @@ {"type":"session","version":0,"id":"33333333-3333-4333-8333-333333333333","createdAt":1789000001000,"cwd":"{{cwd}}","parentSession":"11111111-1111-4111-8111-111111111111","origin":"subagent","delegationDepth":1} -{"type":"subagent/descriptor","data":{"version":2,"mode":"continuable","provider":"spawn","label":"Reply with CHILD_OK","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"continuable","provider":"spawn","label":"Reply with CHILD_OK","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} {"type":"session/end-seed","data":{}} {"type":"sandbox/mode","data":{"mode":"danger-full-access","source":"delegation"}} {"type":"approval/policy","data":{"policy":"never","source":"delegation"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-list-agents/tool-schemas.1.expected.json b/examples/acp-agent/tests/snapshots/subagent-list-agents/tool-schemas.1.expected.json index 38f4eae1ad..62937be9b1 100644 --- a/examples/acp-agent/tests/snapshots/subagent-list-agents/tool-schemas.1.expected.json +++ b/examples/acp-agent/tests/snapshots/subagent-list-agents/tool-schemas.1.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -425,7 +442,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -437,6 +454,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/subagent-max-tokens-partial/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-max-tokens-partial/session.1.jsonl index 4de198e6f6..def81f79a2 100644 --- a/examples/acp-agent/tests/snapshots/subagent-max-tokens-partial/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-max-tokens-partial/session.1.jsonl @@ -5,7 +5,7 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Write the words 'partial one', call todo_write once, then keep going until you are cut off."}],"source":{"kind":"user"},"role":"user","id":"dbf0670a-79cc-4e2c-a298-c4d804e6fe61"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Truncated child"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Truncated child"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Write the words 'partial one', call todo_write once, then keep going until you are cut off."}],"source":{"kind":"user"},"role":"user","id":"dbf0670a-79cc-4e2c-a298-c4d804e6fe61"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"885ea744-63dd-4198-95be-267b9db94a57"},"surfaceOp":"append"} diff --git a/examples/acp-agent/tests/snapshots/subagent-mixed/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-mixed/session.1.jsonl index b64a6cfdfe..3dbae741e9 100644 --- a/examples/acp-agent/tests/snapshots/subagent-mixed/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-mixed/session.1.jsonl @@ -5,10 +5,10 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"73ce401a-faaf-408a-879e-7485380d537d"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Reply ALPHA only"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Reply ALPHA only"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"73ce401a-faaf-408a-879e-7485380d537d"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"f216ca0e-6dcc-4ab3-9cdb-fe38d3dacca2"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"507aa273-ce20-4aaa-9a35-abaae2a5b1cf"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly the word","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-mixed/session.2.jsonl b/examples/acp-agent/tests/snapshots/subagent-mixed/session.2.jsonl index e7de6e1e7d..c623192004 100644 --- a/examples/acp-agent/tests/snapshots/subagent-mixed/session.2.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-mixed/session.2.jsonl @@ -28,10 +28,10 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"What is the project codeword mentioned earlier in this conversation? Reply with exactly that one word and nothing else."}],"source":{"kind":"user"},"role":"user","id":"86e9f144-764f-460d-b72b-262cffe43d77"}]}} {"type":"turn/start","data":{"turn":2}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"fork","label":"Recall project codeword"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"fork","label":"Recall project codeword"}} {"type":"step/start","data":{"turn":2,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"What is the project codeword mentioned earlier in this conversation? Reply with exactly that one word and nothing else."}],"source":{"kind":"user"},"role":"user","id":"86e9f144-764f-460d-b72b-262cffe43d77"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"ac4f4d97-639d-4ad0-a513-219a58355531"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"cf2e06ce-6ea9-451a-bb75-46e59c7a78be"},"surfaceOp":"append"} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"resume"}} {"type":"assistant/chunk","data":{"turn":2,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"reasoning"}}} {"type":"reasoning-chunks","data":{"turn":2,"step":1,"index":0,"dt":[0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"texts":["The"," user"," is"," asking"," me"," to"," recall"," the"," project"," cod","ew","ord"," that"," was"," mentioned"," earlier"," in"," the"," conversation","."," I"," was"," told"," to"," remember"," it",":"," SA","FF","RON","."]}} diff --git a/examples/acp-agent/tests/snapshots/subagent-multi/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-multi/session.1.jsonl index e1b186cd1f..dd6140142a 100644 --- a/examples/acp-agent/tests/snapshots/subagent-multi/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-multi/session.1.jsonl @@ -5,10 +5,10 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"a287f842-f6f2-4a17-ab4c-820e41f498d5"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Return ALPHA only"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Return ALPHA only"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"a287f842-f6f2-4a17-ab4c-820e41f498d5"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"47cdc6a0-a8c8-4842-964a-ad4bc97dc76a"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"ed6eaae0-f071-44ea-9d95-d68185f87194"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly the word","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-multi/session.2.jsonl b/examples/acp-agent/tests/snapshots/subagent-multi/session.2.jsonl index f8cb2e5924..4269894bfb 100644 --- a/examples/acp-agent/tests/snapshots/subagent-multi/session.2.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-multi/session.2.jsonl @@ -5,10 +5,10 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word BETA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"53f6419d-8ddc-4eee-8803-5b68411336f9"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Return BETA only"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Return BETA only"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word BETA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"53f6419d-8ddc-4eee-8803-5b68411336f9"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"f9132345-93c9-40c0-b489-5916bbca96bc"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"de519157-85ec-4e58-9d05-07b469aab403"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly the word","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-parallel/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-parallel/session.1.jsonl index 14b644dc33..1d1f3f1372 100644 --- a/examples/acp-agent/tests/snapshots/subagent-parallel/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-parallel/session.1.jsonl @@ -2,19 +2,19 @@ {"type":"sandbox/mode","data":{"mode":"danger-full-access","source":"delegation"}} {"type":"approval/policy","data":{"policy":"never","source":"delegation"}} {"type":"permission/preset","data":{"preset":"danger-full-access"}} -{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"fadafbc9-263b-4169-82c6-a39868629377"}]}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"e7e63c63-ff17-4f1b-a375-9aba4b477b44"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Say the word ALPHA"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Say the word ALPHA"}} {"type":"step/start","data":{"turn":1,"step":1}} -{"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"fadafbc9-263b-4169-82c6-a39868629377"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"1d6d2982-78f7-49b9-b32d-0eb465d672b1"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"e7e63c63-ff17-4f1b-a375-9aba4b477b44"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"46bdee11-0be5-4a62-a41d-08915b210451"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly the word","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"ALPHA"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} -{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"text","text":"ALPHA"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"9ccb6b64-4dfb-47a2-9967-13ab05483998"}},"sourceEventSeqs":[13,14,15],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"text","text":"ALPHA"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"e195c568-4ea2-4a14-a27c-3ab43d8000b0"}},"sourceEventSeqs":[13,14,15],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":1}} {"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} diff --git a/examples/acp-agent/tests/snapshots/subagent-parallel/session.2.jsonl b/examples/acp-agent/tests/snapshots/subagent-parallel/session.2.jsonl index 5f298c76c5..e708b437fe 100644 --- a/examples/acp-agent/tests/snapshots/subagent-parallel/session.2.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-parallel/session.2.jsonl @@ -2,19 +2,19 @@ {"type":"sandbox/mode","data":{"mode":"danger-full-access","source":"delegation"}} {"type":"approval/policy","data":{"policy":"never","source":"delegation"}} {"type":"permission/preset","data":{"preset":"danger-full-access"}} -{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"dc34a17f-fb30-4afe-a11f-a0d8a1d51658"}]}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"8c91be41-04b6-4c83-a3a6-95e323c807de"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Say the word ALPHA"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Say the word ALPHA"}} {"type":"step/start","data":{"turn":1,"step":1}} -{"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"dc34a17f-fb30-4afe-a11f-a0d8a1d51658"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"12a26f3d-f11e-4de4-8bed-d997590d21e0"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word ALPHA and nothing else."}],"source":{"kind":"user"},"role":"user","id":"8c91be41-04b6-4c83-a3a6-95e323c807de"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"591f5521-ef20-4f12-be4a-420489c9355b"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly the word","messageSeqs":[8],"source":{"kind":"fallback"}}} {"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"ALPHA"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} -{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"text","text":"ALPHA"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"d368f9a5-7d0a-46f0-a7d8-10e1fafa1e74"}},"sourceEventSeqs":[13,14,15],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"text","text":"ALPHA"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"d38d405b-30c9-46b4-a165-78ae723f172e"}},"sourceEventSeqs":[13,14,15],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":1}} {"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} diff --git a/examples/acp-agent/tests/snapshots/subagent-report/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-report/session.1.jsonl index dd28fc243b..e5c3fdaf72 100644 --- a/examples/acp-agent/tests/snapshots/subagent-report/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-report/session.1.jsonl @@ -1,5 +1,5 @@ {"type":"session","version":0,"id":"33333333-3333-4333-8333-333333333333","createdAt":1789000001000,"cwd":"{{cwd}}","parentSession":"11111111-1111-4111-8111-111111111111","origin":"subagent","delegationDepth":1} -{"type":"subagent/descriptor","data":{"version":2,"mode":"continuable","provider":"spawn","label":"Report a finding","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"continuable","provider":"spawn","label":"Report a finding","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} {"type":"session/end-seed","data":{}} {"type":"sandbox/mode","data":{"mode":"danger-full-access","source":"delegation"}} {"type":"approval/policy","data":{"policy":"never","source":"delegation"}} diff --git a/examples/acp-agent/tests/snapshots/subagent-report/tool-schemas.1.expected.json b/examples/acp-agent/tests/snapshots/subagent-report/tool-schemas.1.expected.json index 38f4eae1ad..62937be9b1 100644 --- a/examples/acp-agent/tests/snapshots/subagent-report/tool-schemas.1.expected.json +++ b/examples/acp-agent/tests/snapshots/subagent-report/tool-schemas.1.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -425,7 +442,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -437,6 +454,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/subagent-spawn-in-process/session.1.jsonl b/examples/acp-agent/tests/snapshots/subagent-spawn-in-process/session.1.jsonl index efe72ff42b..a70c330e78 100644 --- a/examples/acp-agent/tests/snapshots/subagent-spawn-in-process/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/subagent-spawn-in-process/session.1.jsonl @@ -5,7 +5,7 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"54ed23d6-e960-4f36-b192-cf06e1618ea6"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Reply with CHILD_OK"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Reply with CHILD_OK"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"54ed23d6-e960-4f36-b192-cf06e1618ea6"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"24630f5a-f790-469f-96a6-cf234ded3759"},"surfaceOp":"append"} diff --git a/examples/acp-agent/tests/snapshots/text-turn/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/text-turn/tool-schemas.expected.json index db3c652d58..0720890967 100644 --- a/examples/acp-agent/tests/snapshots/text-turn/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/text-turn/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -409,7 +426,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -421,6 +438,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/web-fetch/tool-schemas.expected.json b/examples/acp-agent/tests/snapshots/web-fetch/tool-schemas.expected.json index b2236d44a3..630a9b086f 100644 --- a/examples/acp-agent/tests/snapshots/web-fetch/tool-schemas.expected.json +++ b/examples/acp-agent/tests/snapshots/web-fetch/tool-schemas.expected.json @@ -260,6 +260,23 @@ } } }, + { + "name": "list_subagent_models", + "description": "Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.", + "parameters": { + "type": "object", + "properties": { + "provider": { + "type": "string", + "description": "Registered LLM provider id. Omit to list providers." + }, + "model": { + "type": "string", + "description": "Exact model id to inspect. Requires provider; omit to list that provider's advertised models." + } + } + } + }, { "name": "ralph", "description": "Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.", @@ -409,7 +426,7 @@ }, { "name": "subagent", - "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.", + "description": "Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.", "parameters": { "type": "object", "properties": { @@ -421,6 +438,18 @@ "type": "string", "description": "The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs." }, + "provider": { + "type": "string", + "description": "LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route." + }, + "model": { + "type": "string", + "description": "Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route." + }, + "reasoning_effort": { + "type": "string", + "description": "Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default." + }, "run_in_background": { "type": "boolean", "description": "Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it." diff --git a/examples/acp-agent/tests/snapshots/workflow-run/session.1.jsonl b/examples/acp-agent/tests/snapshots/workflow-run/session.1.jsonl index a42b3cf01d..fce39e713c 100644 --- a/examples/acp-agent/tests/snapshots/workflow-run/session.1.jsonl +++ b/examples/acp-agent/tests/snapshots/workflow-run/session.1.jsonl @@ -5,7 +5,7 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly the word WF_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"f0f46771-663a-494a-8d40-6866a5bbe7c9"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly the word WF_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"f0f46771-663a-494a-8d40-6866a5bbe7c9"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"12bbd4dd-4040-4cc7-8acf-e526144f1ee5"},"surfaceOp":"append"} diff --git a/examples/headless-agent/cordis.yml b/examples/headless-agent/cordis.yml index fa037ab935..759a252b4e 100644 --- a/examples/headless-agent/cordis.yml +++ b/examples/headless-agent/cordis.yml @@ -124,13 +124,17 @@ config: provider: spawn toolName: subagent + enableModelSelection: true backgroundMode: continuable maxDepth: 1 -# Fork stays one-shot because a continuable child's `report` tool and prompt -# section precede the inherited history a fork reuses; `run_in_background` is off -# as an explicit foreground-only choice even though agent-spine-demo mounts the -# generic Job runtime. See .agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md. +# Fork omits model selection so provider/model stay equal to the parent and the +# inherited history remains eligible for KV Cache reuse. It stays one-shot because +# a continuable child's `report` tool and prompt section precede that history and +# invalidate the same prefix. `run_in_background` is off as an explicit +# foreground-only choice even though agent-spine-demo mounts the generic Job runtime. +# See .agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md +# and .agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md. - id: tool-subagent-fork name: '@deepseek-ai/dsh-tool-subagent' config: diff --git a/examples/headless-agent/tests/snapshots/advanced-toolchain/session.1.jsonl b/examples/headless-agent/tests/snapshots/advanced-toolchain/session.1.jsonl index 1287de6339..6455b03e27 100644 --- a/examples/headless-agent/tests/snapshots/advanced-toolchain/session.1.jsonl +++ b/examples/headless-agent/tests/snapshots/advanced-toolchain/session.1.jsonl @@ -1,19 +1,19 @@ {"type":"session","version":0,"id":"22222222-2222-4222-8222-222222222222","createdAt":1783950001000,"cwd":"{{cwd}}","parentSession":"11111111-1111-4111-8111-111111111111","origin":"subagent","delegationDepth":1} -{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly DIRECT_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"c66e310e-2597-4d01-85c8-2d70a9d831c0"}]}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly DIRECT_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"2324a504-7992-4dd5-b1a4-22783caf793c"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Check direct child"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Check direct child"}} {"type":"step/start","data":{"turn":1,"step":1}} -{"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly DIRECT_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"c66e310e-2597-4d01-85c8-2d70a9d831c0"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"fd0a0587-8df4-46b2-809a-317346a4c0f4"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly DIRECT_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"2324a504-7992-4dd5-b1a4-22783caf793c"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"9b642d5e-ecc8-4c09-8f98-fd067b62ae63"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly DIRECT_CHILD_OK and","messageSeqs":[5],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model. Your working directory is {{cwd}}.\n\nVerify your work by running the code or tests. Keep answers brief and factual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\n# Dynamic Cordis Plugins\n\nDynamic Cordis plugins temporarily extend the current DSH process. A Plugin uses apply(ctx) to consume Services, listen to Events, provide Services, register model Tools, or register browser UI in Slots.\n\n- Plugin and Package definitions exist only in the current process. define itself does not modify repository source, configuration, or disk, and definitions do not survive a process restart.\n- The restricted execution environment prevents accidental misuse; it is not a security boundary for malicious code. Services obtained by dynamic code connect to the real runtime.\n\n## Make the user-facing plan clear first\n\n- Dynamic Cordis Plugins are one available implementation mechanism, not the default for every request. Consider whether one could help only when the user intends to design or create something, or when a temporary interface could materially aid the current work. The presence of these instructions or Tools, and discussion of Cordis itself, do not make a request a dynamic-Plugin task.\n- When Cordis is a plausible fit, infer the intended work target and lifetime from the request and conversation. Use it only when the outcome belongs to the current running harness and should be delivered as a temporary runtime extension. If that distinction is materially ambiguous, ask at most one concise question about the intended result or lifetime. Otherwise proceed with the matching workflow; do not require the user to know or choose Cordis as an implementation mechanism.\n- Once a dynamic Plugin is appropriate, decide whether the task creates a new Plugin or modifies the Plugin named by the user with @pluginId. Proceed directly when the goal is clear; do not ask for repeated confirmation.\n- Choose Host, Client, or both from the requested outcome. Do not propose a Client/browser UI when the task does not need visible page behavior, and do not avoid Client when the requested outcome is visual, interactive, or depends on page state. Host versus Client is an implementation choice; do not make the user choose it.\n- When a design direction or a potentially useful interface would materially affect the result, ask at most one concise outcome or creative-preference question and offer a few candidate directions. Otherwise proceed directly; do not conduct a multi-round interview or a complex questionnaire.\n- cordis_define only defines and presents code; it does not run it. After definition, explain the pluginId and packageId returned by the Host and whether the next step is a run or update.\n- cordis_run may require user approval. When it returns awaiting-approval, explain that the user must allow or reject it in the UI. Do not wait, retry, or claim that it is running.\n- When it returns starting, explain that the request has entered the asynchronous flow and the Client is still activating. starting does not mean success. Wait for the system to report the final result through steering context.\n- Do not request approval again after the user rejects it. After a technical failure, fix the same Plugin from its diagnostics; do not silently create a replacement Plugin.\n\n## Recommended workflow and Tools\n\nBefore creating, modifying, or repairing a Plugin, load the cordis-plugin-development Skill. The Skill provides requirement navigation, capability composition, complete examples, and troubleshooting. Treat Inspect Provider results as the source of truth for exact APIs.\n\n1. cordis_inspect_list: discover the current Host and Client Providers and their read-only query methods.\n2. cordis_inspect_query: use the returned platform, provider, method, and schema to query exact Service, Event, Builtin, Slot, Theme token, or Tool information.\n3. cordis_inspect_self: inspect the current Session's Plugins, Packages, version pointers, source, and diagnostics. Source is returned only when both pluginId and packageId are specified.\n4. cordis_define: create the first Package for a new Plugin or append an immutable Package to an existing Plugin. It defines code but does not run it.\n5. cordis_run: activate an exact Package. Use run for the first activation, restarting current, or rollback; use update to switch versions.\n6. cordis_stop: remove the current Run and pending approval request while retaining definitions, grants, and version pointers.\n7. cordis_undefine: permanently stop and delete a Plugin and all of its Packages. Use it only after confirming that the user no longer needs them.\n\n- Inspect and Catalog data only confirm capabilities, names, signatures, types, and registration protocols before code is written; they do not replace business APIs.\n- Query Service.listService and Event.listEvents without input to choose from their compact signature directories, then query the exact service or event before using it. Exact queries return the structured contract and only its referenced types.\n- At runtime, a Plugin must call real Services or listen to real Events. Do not cache, display, or depend on Inspect results as business data.\n\n## Identity, versions, and approval\n\n- pluginId identifies a Plugin that can be modified over time. For a new Plugin, submit only a semantic idPrefix of 3–6 lowercase English letters; the Host allocates the final ID.\n- packageId identifies one immutable Host/Client source version under a Plugin. To change code, define a new Package; never overwrite an old version.\n- pluginRunId identifies one activation attempt and connects its approval, Host/Client loading, private RPC, Run card, and errors.\n- currentPackageId is the most recent fully successful Package. Stopping, starting an update, or failing an update does not clear it.\n- nextPackageId is the target awaiting approval, being attempted, awaiting Client activation, or most recently failed.\n- A single check mark authorizes only the current Package; double check marks authorize future versions of the same Plugin. A grant remains in effect after a technical failure.\n- An update stops the old Run before starting the target Package. Failure does not automatically restart the old version; retry next with update or roll back to current with run.\n\nWhen the user enters @pluginId, the system injects identity, the default base Package, version pointers, and runtime status, but not source code:\n\n1. Call cordis_inspect_self(pluginId, packageId) to read the target source.\n2. Use cordis_define in existing mode to append a Package to the same Plugin.\n3. Call cordis_run in run or update mode according to the version relationship.\n\nNever silently create another Plugin for @pluginId. If the reference is unavailable because it was removed, belongs to another Session, or was lost on process restart, tell the user directly.\n\n## High-frequency errors that must be avoided\n\n### Services: ctx.get and inject\n\n- Read an optional Service with ctx.get('serviceName') by default and handle undefined.\n- Declare inject: ['serviceName'] on the returned Plugin object only when the Service is a hard dependency and the Plugin must enter waiting until Cordis reactivates it after the Service appears.\n- Read ctx.serviceName only after declaring that Service in inject. Never access an undeclared Service as a ctx property.\n\n```js\nreturn {\n inject: ['requiredService'],\n apply(ctx) {\n ctx.requiredService.someMethod()\n const optionalService = ctx.get('optionalService')\n if (optionalService !== undefined) optionalService.someMethod()\n },\n}\n```\n\n### Code: use plain JavaScript only\n\n- Host and Client code is not transformed by TypeScript, JSX, or a bundler.\n- Do not use TypeScript types, as, decorators, import, require, or JSX.\n- Client React code must use React.createElement(...); never write .\n- Do not assume that process, Buffer, window, document, fetch, native timers, or any other global is available. Query the corresponding platform's Builtins and Services first.\n\n### Data: do not serialize live data\n\n- Services, Events, Slots, Sessions, and their derived Cordis/DSH objects are internal live data, not ordinary JSON that can be dumped.\n- Do not apply JSON.stringify, structuredClone, recursive enumeration, full copying, or whole-object display to live data.\n- Read only the leaf fields required by the task, then construct the smallest owned data object without Host references.\n\n### Lifecycle: every side effect must be reversible\n\n- Services, Events, Tools, handlers, timers, Slots, styles, and theme overrides must all belong to the current Fiber.\n- Use ctx.effect(), ctx.on(), or official APIs that return a disposer so stop, update, or undefine removes every side effect.\n- The cordis-plugin-development Skill contains complete timer, Waterfall, Slot, theme, Tool, RPC, and React examples and troubleshooting guidance.\n\n## Host and Client\n\n- Host runs in the DSH Node.js process and is appropriate for files, networking, commands, Agent/Session access, Host Events, Services, model Tools, and JSON methods callable by the Client.\n- Client runs in the browser page and is appropriate for themes, layout, current page state, Tool cards, and Slot UI.\n- Host and Client communicate through Package-private JSON methods: Host uses harness.handle(method, handler), and Client uses host.call(method, args). The direction is Client→Host, and only lossless JSON may cross it.\n- Client UI must be registered in a queried Slot; apply() cannot directly return a React Element. Query Slots.listSubTree without root to choose from the compact purpose/topology tree, then query the exact root for its full registration contract and props before writing code.\n- See the Skill and Inspect Providers for Run-specific panels and exact Slot registration patterns.\n\n## Asynchronous results and recovery\n\n- Do not wait inside a Tool for approval or browser work that can happen only after the current turn ends.\n- Asynchronous success, rejection, and runtime errors update Run state and notify you through steering context.\n- After a technical failure, use cordis_inspect_self to read the exact Package source and its message/stack. Define a corrected Package under the same Plugin and retry autonomously.\n- Use the cordis-plugin-development Skill for other failure causes, repair procedures, and complete extension patterns.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.\n\n## Writing code for run_code\n\n`run_code` takes two required arguments: `code` — the body of an async TypeScript function (erasable syntax only — no `enum` or namespaces; type annotations are advisory, the code runs type-stripped) — and `description`, a short summary of what the program does. Inside the program:\n\n- Call tools as `await tools.name(args)` — quoted access for exotic names: `tools[\"my-tool\"](args)`. Every call resolves to the tool's typed canonical JSON value. Tool arguments must be lossless JSON.\n- A FAILED tool call rejects with `ToolCallError`, whose `toolName` identifies the failed tool and whose `message` is human-readable — `try/catch` it to handle and continue.\n- Independent read-only calls MAY overlap under `Promise.all` (safe calls run concurrently; mutating calls run alone, in submission order). Sequence dependent work with `await`.\n- Emit results with `return` and/or `console.log(...)`. ONLY what you print or return comes back to you — intermediate tool results never enter the conversation, so extract just what you need.\n\nThe available tools:\n\n```ts\ntype JsonValue = null | boolean | number | string | JsonValue[] | { [key: string]: JsonValue }\n\ninterface ToolArgsMap {\n /** Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`. */\n bash: {\n /** The bash command to execute. */\n command: string;\n /** Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\". */\n description: string;\n /** Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry. */\n timeoutMs?: number;\n /** Working directory for this command. Defaults to the session workspace; a relative path is resolved against it. */\n workdir?: string;\n /** Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies. */\n run_in_background?: boolean;\n } & Record;\n /** Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs. */\n cordis_define: {\n plugin: {\n kind: \"new\";\n /** Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix. */\n idPrefix: string;\n } | {\n kind: \"existing\";\n /** Exact ID of an existing Plugin; the new Package is appended to that instance. */\n pluginId: string;\n };\n /** Short, readable Package name. */\n name: string;\n /** One-sentence, user-facing description of the Package purpose. */\n purpose: string;\n code: {\n /** Plain JavaScript function body that returns the Host-half Cordis Plugin. */\n host?: string;\n /** Plain JavaScript function body that returns the browser Client-half Cordis Plugin. */\n client?: string;\n };\n } & Record;\n /** List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call. */\n cordis_inspect_list: Record;\n /** Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props. */\n cordis_inspect_query: {\n /** Runtime platform that owns the Provider. */\n platform: \"host\" | \"client\";\n /** Exact Provider ID returned by cordis_inspect_list. */\n provider: string;\n /** Exact method name declared by the Provider manifest. */\n method: string;\n /** Optional query input; it must satisfy the method input schema. */\n input?: JsonValue;\n } & Record;\n /** Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers. */\n cordis_inspect_self: {\n /** Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin. */\n pluginId?: string;\n /** Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned. */\n packageId?: string;\n } & Record;\n /** Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it. */\n cordis_run: {\n /** Stable Plugin ID returned by cordis_define. */\n pluginId: string;\n /** Exact immutable Package ID to activate under that Plugin. */\n packageId: string;\n /** Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package. */\n mode: \"run\" | \"update\";\n } & Record;\n /** Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal. */\n cordis_stop: {\n /** Stable dynamic Plugin ID to stop. */\n pluginId: string;\n } & Record;\n /** Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead. */\n cordis_undefine: {\n /** Stable dynamic Plugin ID to remove permanently. */\n pluginId: string;\n } & Record;\n /** Edit an existing UTF-8 text file by replacing literal text. */\n edit: {\n /** Path to edit, resolved by the filesystem backend. */\n file_path: string;\n /** Literal text to replace. Must match exactly. */\n old_string: string;\n /** Literal replacement text. Use an empty string to delete the match. */\n new_string: string;\n /** Replace all matches. Defaults to false; when false, old_string must appear exactly once. */\n replace_all?: boolean;\n } & Record;\n /** Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op. */\n interrupt_agent: {\n /** The agent id of the running agent to interrupt. */\n agent_id: string;\n } & Record;\n /** Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops. */\n job_kill: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Optional short reason, recorded in the log and forwarded to the job. */\n reason?: string;\n } & Record;\n /** List your background jobs (running and finished) with their ids, kinds, and statuses. */\n job_list: Record;\n /** Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap. */\n job_output: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive. */\n wait?: boolean;\n /** Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum. */\n timeout_ms?: number;\n } & Record;\n /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */\n ralph: {\n /** The immutable completion objective for every fresh Ralph round. */\n objective: string;\n /** Optional positive safe-integer round cap, bounded by the deployment ceiling. */\n maxRounds?: number;\n } & Record;\n /** Read a UTF-8 text file and return line-numbered content. */\n read: {\n /** Path to read, resolved by the filesystem backend. */\n file_path: string;\n /** 1-based first line to return. Defaults to 1. */\n offset?: number;\n /** Maximum number of lines to return. Defaults to 2000. */\n limit?: number;\n } & Record;\n /** Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered. */\n send_message: {\n /** The subagent id returned when the background subagent was started. */\n subagent_id: string;\n /** The message to deliver to the subagent. */\n message: string;\n } & Record;\n /** Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill. */\n skill: {\n /** The exact skill name from the available skills list. */\n name: string;\n } & Record;\n /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. */\n subagent: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */\n prompt: string;\n /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */\n run_in_background?: boolean;\n } & Record;\n /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result. */\n subagent_fork: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new. */\n prompt: string;\n } & Record;\n /** Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished). */\n todo_write: {\n /** The COMPLETE task list, replacing any previous list. */\n todos: ({\n /** What the task is — a short imperative line. */\n content: string;\n /** pending (not started) | in_progress (now) | completed (done). */\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n } & Record;\n /** Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn. The workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result. Script-body hooks: - `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly. - `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages. - `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`. - `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim. Misused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`. Constraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes. */\n workflow: {\n /** The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `). */\n script: string;\n /** The workflow identity block (plain JSON — never code). */\n meta: {\n /** Short kebab-case workflow name. */\n name: string;\n /** One-line description of what the workflow does. */\n description: string;\n /** Optional guidance on when this workflow applies. */\n whenToUse?: string;\n /** Optional phase declarations matched by phase() calls. */\n phases?: ({\n /** The phase title phase() calls match by exact string. */\n title: string;\n /** Optional one-line description of the phase. */\n detail?: string;\n /** Optional provider override this phase is expected to use. */\n provider?: string;\n /** Optional model override this phase is expected to use. */\n model?: string;\n } & Record)[];\n } & Record;\n /** Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}). */\n args?: Record;\n } & Record;\n /** Create or fully replace a UTF-8 text file. */\n write: {\n /** Path to write, resolved by the filesystem backend. */\n file_path: string;\n /** Full UTF-8 text content to write. */\n content: string;\n } & Record;\n}\n\ninterface ToolOutputMap {\n bash: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"foreground\";\n exitCode: number | null;\n signal: string | null;\n timedOut: boolean;\n aborted: boolean;\n timeoutMs: number;\n stdout: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n stderr: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n sandbox?: {\n mode: string;\n denied: boolean;\n enforcement?: string;\n runnerFailed?: boolean;\n };\n };\n cordis_define: {\n pluginId: string;\n packageId: string;\n name: string;\n purpose: string;\n hasHostHalf: boolean;\n hasClientHalf: boolean;\n };\n cordis_inspect_list: JsonValue;\n cordis_inspect_query: JsonValue;\n cordis_inspect_self: JsonValue;\n cordis_run: JsonValue;\n cordis_stop: {\n pluginId: string;\n };\n cordis_undefine: {\n pluginId: string;\n wasRunning: boolean;\n };\n edit: {\n path: string;\n before: string;\n after: string;\n };\n interrupt_agent: {\n accepted: boolean;\n };\n job_kill: {\n outcome: \"cancellation-requested\" | \"already-finished\";\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n job_list: ({\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n })[];\n job_output: {\n text: string;\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n ralph: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n read: {\n path: string;\n offset: number;\n lines: {\n number: number;\n text: string;\n }[];\n totalLines: number;\n };\n send_message: {\n messageId: string;\n };\n skill: {\n name: string;\n provider: string;\n resourceBase?: {\n kind: \"directory\";\n path: string;\n } | {\n kind: \"url\";\n url: string;\n } | {\n kind: \"opaque\";\n description: string;\n };\n content: string;\n };\n subagent: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n subagent_fork: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n todo_write: {\n todos: ({\n content: string;\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n counts: {\n pending: number;\n inProgress: number;\n completed: number;\n };\n };\n workflow: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n write: {\n path: string;\n operation: \"create\" | \"update\";\n before: string | null;\n after: string;\n };\n}\n\ntype ToolName = keyof ToolOutputMap\n\ndeclare class ToolCallError extends Error {\n readonly name: \"ToolCallError\";\n readonly toolName: ToolName;\n}\n\ndeclare const tools: {\n [K in ToolName]: (args: ToolArgsMap[K]) => Promise;\n}\n```","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"cordis_define","description":"Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs.","parameters":{"type":"object","properties":{"plugin":{"oneOf":[{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"new"},"idPrefix":{"type":"string","description":"Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix."}},"required":["kind","idPrefix"]},{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"existing"},"pluginId":{"type":"string","description":"Exact ID of an existing Plugin; the new Package is appended to that instance."}},"required":["kind","pluginId"]}]},"name":{"type":"string","description":"Short, readable Package name."},"purpose":{"type":"string","description":"One-sentence, user-facing description of the Package purpose."},"code":{"type":"object","additionalProperties":false,"properties":{"host":{"type":"string","description":"Plain JavaScript function body that returns the Host-half Cordis Plugin."},"client":{"type":"string","description":"Plain JavaScript function body that returns the browser Client-half Cordis Plugin."}}}},"required":["plugin","name","purpose","code"]}},{"name":"cordis_inspect_list","description":"List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call.","parameters":{"type":"object","properties":{}}},{"name":"cordis_inspect_query","description":"Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props.","parameters":{"type":"object","properties":{"platform":{"type":"string","description":"Runtime platform that owns the Provider.","enum":["host","client"]},"provider":{"type":"string","description":"Exact Provider ID returned by cordis_inspect_list."},"method":{"type":"string","description":"Exact method name declared by the Provider manifest."},"input":{"description":"Optional query input; it must satisfy the method input schema."}},"required":["platform","provider","method"]}},{"name":"cordis_inspect_self","description":"Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin."},"packageId":{"type":"string","description":"Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned."}}}},{"name":"cordis_run","description":"Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define."},"packageId":{"type":"string","description":"Exact immutable Package ID to activate under that Plugin."},"mode":{"type":"string","description":"Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package.","enum":["run","update"]}},"required":["pluginId","packageId","mode"]}},{"name":"cordis_stop","description":"Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to stop."}},"required":["pluginId"]}},{"name":"cordis_undefine","description":"Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to remove permanently."}},"required":["pluginId"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"run_code","description":"Execute a TypeScript program against the available tools. Takes two required arguments: `code`, the BODY of an async function (erasable syntax only; top-level `await` and `return` work), and `description`, a short summary of what the program does. Call tools as `await tools.name(args)` per the declarations in the system prompt. Only what you print or return comes back — curate it.","parameters":{"type":"object","properties":{"code":{"type":"string","description":"The program: the body of an async TypeScript function."},"description":{"type":"string","description":"Clear, concise description of what this program does in active voice, 5-10 words (shown in the UI). Examples: \"Count TODO markers across packages\"; \"Read failing test and its fixture\"; \"Rename config key in every cordis.yml\"."}},"required":["code","description"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model. Your working directory is {{cwd}}.\n\nVerify your work by running the code or tests. Keep answers brief and factual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\n# Dynamic Cordis Plugins\n\nDynamic Cordis plugins temporarily extend the current DSH process. A Plugin uses apply(ctx) to consume Services, listen to Events, provide Services, register model Tools, or register browser UI in Slots.\n\n- Plugin and Package definitions exist only in the current process. define itself does not modify repository source, configuration, or disk, and definitions do not survive a process restart.\n- The restricted execution environment prevents accidental misuse; it is not a security boundary for malicious code. Services obtained by dynamic code connect to the real runtime.\n\n## Make the user-facing plan clear first\n\n- Dynamic Cordis Plugins are one available implementation mechanism, not the default for every request. Consider whether one could help only when the user intends to design or create something, or when a temporary interface could materially aid the current work. The presence of these instructions or Tools, and discussion of Cordis itself, do not make a request a dynamic-Plugin task.\n- When Cordis is a plausible fit, infer the intended work target and lifetime from the request and conversation. Use it only when the outcome belongs to the current running harness and should be delivered as a temporary runtime extension. If that distinction is materially ambiguous, ask at most one concise question about the intended result or lifetime. Otherwise proceed with the matching workflow; do not require the user to know or choose Cordis as an implementation mechanism.\n- Once a dynamic Plugin is appropriate, decide whether the task creates a new Plugin or modifies the Plugin named by the user with @pluginId. Proceed directly when the goal is clear; do not ask for repeated confirmation.\n- Choose Host, Client, or both from the requested outcome. Do not propose a Client/browser UI when the task does not need visible page behavior, and do not avoid Client when the requested outcome is visual, interactive, or depends on page state. Host versus Client is an implementation choice; do not make the user choose it.\n- When a design direction or a potentially useful interface would materially affect the result, ask at most one concise outcome or creative-preference question and offer a few candidate directions. Otherwise proceed directly; do not conduct a multi-round interview or a complex questionnaire.\n- cordis_define only defines and presents code; it does not run it. After definition, explain the pluginId and packageId returned by the Host and whether the next step is a run or update.\n- cordis_run may require user approval. When it returns awaiting-approval, explain that the user must allow or reject it in the UI. Do not wait, retry, or claim that it is running.\n- When it returns starting, explain that the request has entered the asynchronous flow and the Client is still activating. starting does not mean success. Wait for the system to report the final result through steering context.\n- Do not request approval again after the user rejects it. After a technical failure, fix the same Plugin from its diagnostics; do not silently create a replacement Plugin.\n\n## Recommended workflow and Tools\n\nBefore creating, modifying, or repairing a Plugin, load the cordis-plugin-development Skill. The Skill provides requirement navigation, capability composition, complete examples, and troubleshooting. Treat Inspect Provider results as the source of truth for exact APIs.\n\n1. cordis_inspect_list: discover the current Host and Client Providers and their read-only query methods.\n2. cordis_inspect_query: use the returned platform, provider, method, and schema to query exact Service, Event, Builtin, Slot, Theme token, or Tool information.\n3. cordis_inspect_self: inspect the current Session's Plugins, Packages, version pointers, source, and diagnostics. Source is returned only when both pluginId and packageId are specified.\n4. cordis_define: create the first Package for a new Plugin or append an immutable Package to an existing Plugin. It defines code but does not run it.\n5. cordis_run: activate an exact Package. Use run for the first activation, restarting current, or rollback; use update to switch versions.\n6. cordis_stop: remove the current Run and pending approval request while retaining definitions, grants, and version pointers.\n7. cordis_undefine: permanently stop and delete a Plugin and all of its Packages. Use it only after confirming that the user no longer needs them.\n\n- Inspect and Catalog data only confirm capabilities, names, signatures, types, and registration protocols before code is written; they do not replace business APIs.\n- Query Service.listService and Event.listEvents without input to choose from their compact signature directories, then query the exact service or event before using it. Exact queries return the structured contract and only its referenced types.\n- At runtime, a Plugin must call real Services or listen to real Events. Do not cache, display, or depend on Inspect results as business data.\n\n## Identity, versions, and approval\n\n- pluginId identifies a Plugin that can be modified over time. For a new Plugin, submit only a semantic idPrefix of 3–6 lowercase English letters; the Host allocates the final ID.\n- packageId identifies one immutable Host/Client source version under a Plugin. To change code, define a new Package; never overwrite an old version.\n- pluginRunId identifies one activation attempt and connects its approval, Host/Client loading, private RPC, Run card, and errors.\n- currentPackageId is the most recent fully successful Package. Stopping, starting an update, or failing an update does not clear it.\n- nextPackageId is the target awaiting approval, being attempted, awaiting Client activation, or most recently failed.\n- A single check mark authorizes only the current Package; double check marks authorize future versions of the same Plugin. A grant remains in effect after a technical failure.\n- An update stops the old Run before starting the target Package. Failure does not automatically restart the old version; retry next with update or roll back to current with run.\n\nWhen the user enters @pluginId, the system injects identity, the default base Package, version pointers, and runtime status, but not source code:\n\n1. Call cordis_inspect_self(pluginId, packageId) to read the target source.\n2. Use cordis_define in existing mode to append a Package to the same Plugin.\n3. Call cordis_run in run or update mode according to the version relationship.\n\nNever silently create another Plugin for @pluginId. If the reference is unavailable because it was removed, belongs to another Session, or was lost on process restart, tell the user directly.\n\n## High-frequency errors that must be avoided\n\n### Services: ctx.get and inject\n\n- Read an optional Service with ctx.get('serviceName') by default and handle undefined.\n- Declare inject: ['serviceName'] on the returned Plugin object only when the Service is a hard dependency and the Plugin must enter waiting until Cordis reactivates it after the Service appears.\n- Read ctx.serviceName only after declaring that Service in inject. Never access an undeclared Service as a ctx property.\n\n```js\nreturn {\n inject: ['requiredService'],\n apply(ctx) {\n ctx.requiredService.someMethod()\n const optionalService = ctx.get('optionalService')\n if (optionalService !== undefined) optionalService.someMethod()\n },\n}\n```\n\n### Code: use plain JavaScript only\n\n- Host and Client code is not transformed by TypeScript, JSX, or a bundler.\n- Do not use TypeScript types, as, decorators, import, require, or JSX.\n- Client React code must use React.createElement(...); never write .\n- Do not assume that process, Buffer, window, document, fetch, native timers, or any other global is available. Query the corresponding platform's Builtins and Services first.\n\n### Data: do not serialize live data\n\n- Services, Events, Slots, Sessions, and their derived Cordis/DSH objects are internal live data, not ordinary JSON that can be dumped.\n- Do not apply JSON.stringify, structuredClone, recursive enumeration, full copying, or whole-object display to live data.\n- Read only the leaf fields required by the task, then construct the smallest owned data object without Host references.\n\n### Lifecycle: every side effect must be reversible\n\n- Services, Events, Tools, handlers, timers, Slots, styles, and theme overrides must all belong to the current Fiber.\n- Use ctx.effect(), ctx.on(), or official APIs that return a disposer so stop, update, or undefine removes every side effect.\n- The cordis-plugin-development Skill contains complete timer, Waterfall, Slot, theme, Tool, RPC, and React examples and troubleshooting guidance.\n\n## Host and Client\n\n- Host runs in the DSH Node.js process and is appropriate for files, networking, commands, Agent/Session access, Host Events, Services, model Tools, and JSON methods callable by the Client.\n- Client runs in the browser page and is appropriate for themes, layout, current page state, Tool cards, and Slot UI.\n- Host and Client communicate through Package-private JSON methods: Host uses harness.handle(method, handler), and Client uses host.call(method, args). The direction is Client→Host, and only lossless JSON may cross it.\n- Client UI must be registered in a queried Slot; apply() cannot directly return a React Element. Query Slots.listSubTree without root to choose from the compact purpose/topology tree, then query the exact root for its full registration contract and props before writing code.\n- See the Skill and Inspect Providers for Run-specific panels and exact Slot registration patterns.\n\n## Asynchronous results and recovery\n\n- Do not wait inside a Tool for approval or browser work that can happen only after the current turn ends.\n- Asynchronous success, rejection, and runtime errors update Run state and notify you through steering context.\n- After a technical failure, use cordis_inspect_self to read the exact Package source and its message/stack. Define a corrected Package under the same Plugin and retry autonomously.\n- Use the cordis-plugin-development Skill for other failure causes, repair procedures, and complete extension patterns.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.\n\n## Writing code for run_code\n\n`run_code` takes two required arguments: `code` — the body of an async TypeScript function (erasable syntax only — no `enum` or namespaces; type annotations are advisory, the code runs type-stripped) — and `description`, a short summary of what the program does. Inside the program:\n\n- Call tools as `await tools.name(args)` — quoted access for exotic names: `tools[\"my-tool\"](args)`. Every call resolves to the tool's typed canonical JSON value. Tool arguments must be lossless JSON.\n- A FAILED tool call rejects with `ToolCallError`, whose `toolName` identifies the failed tool and whose `message` is human-readable — `try/catch` it to handle and continue.\n- Independent read-only calls MAY overlap under `Promise.all` (safe calls run concurrently; mutating calls run alone, in submission order). Sequence dependent work with `await`.\n- Emit results with `return` and/or `console.log(...)`. Only what you print or return is program output. A successful tool result containing an image is attached after the run so you can inspect it on the next step; every other intermediate result stays out of the conversation, so extract just what you need.\n\nThe available tools:\n\n```ts\ntype JsonValue = null | boolean | number | string | JsonValue[] | { [key: string]: JsonValue }\n\ninterface ToolArgsMap {\n /** Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`. */\n bash: {\n /** The bash command to execute. */\n command: string;\n /** Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\". */\n description: string;\n /** Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry. */\n timeoutMs?: number;\n /** Working directory for this command. Defaults to the session workspace; a relative path is resolved against it. */\n workdir?: string;\n /** Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies. */\n run_in_background?: boolean;\n } & Record;\n /** Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs. */\n cordis_define: {\n plugin: {\n kind: \"new\";\n /** Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix. */\n idPrefix: string;\n } | {\n kind: \"existing\";\n /** Exact ID of an existing Plugin; the new Package is appended to that instance. */\n pluginId: string;\n };\n /** Short, readable Package name. */\n name: string;\n /** One-sentence, user-facing description of the Package purpose. */\n purpose: string;\n code: {\n /** Plain JavaScript function body that returns the Host-half Cordis Plugin. */\n host?: string;\n /** Plain JavaScript function body that returns the browser Client-half Cordis Plugin. */\n client?: string;\n };\n } & Record;\n /** List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call. */\n cordis_inspect_list: Record;\n /** Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props. */\n cordis_inspect_query: {\n /** Runtime platform that owns the Provider. */\n platform: \"host\" | \"client\";\n /** Exact Provider ID returned by cordis_inspect_list. */\n provider: string;\n /** Exact method name declared by the Provider manifest. */\n method: string;\n /** Optional query input; it must satisfy the method input schema. */\n input?: JsonValue;\n } & Record;\n /** Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers. */\n cordis_inspect_self: {\n /** Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin. */\n pluginId?: string;\n /** Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned. */\n packageId?: string;\n } & Record;\n /** Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it. */\n cordis_run: {\n /** Stable Plugin ID returned by cordis_define. */\n pluginId: string;\n /** Exact immutable Package ID to activate under that Plugin. */\n packageId: string;\n /** Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package. */\n mode: \"run\" | \"update\";\n } & Record;\n /** Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal. */\n cordis_stop: {\n /** Stable dynamic Plugin ID to stop. */\n pluginId: string;\n } & Record;\n /** Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead. */\n cordis_undefine: {\n /** Stable dynamic Plugin ID to remove permanently. */\n pluginId: string;\n } & Record;\n /** Edit an existing UTF-8 text file by replacing literal text. */\n edit: {\n /** Path to edit, resolved by the filesystem backend. */\n file_path: string;\n /** Literal text to replace. Must match exactly. */\n old_string: string;\n /** Literal replacement text. Use an empty string to delete the match. */\n new_string: string;\n /** Replace all matches. Defaults to false; when false, old_string must appear exactly once. */\n replace_all?: boolean;\n } & Record;\n /** Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op. */\n interrupt_agent: {\n /** The agent id of the running agent to interrupt. */\n agent_id: string;\n } & Record;\n /** Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops. */\n job_kill: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Optional short reason, recorded in the log and forwarded to the job. */\n reason?: string;\n } & Record;\n /** List your background jobs (running and finished) with their ids, kinds, and statuses. */\n job_list: Record;\n /** Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap. */\n job_output: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive. */\n wait?: boolean;\n /** Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum. */\n timeout_ms?: number;\n } & Record;\n /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */\n list_subagent_models: {\n /** Registered LLM provider id. Omit to list providers. */\n provider?: string;\n /** Exact model id to inspect. Requires provider; omit to list that provider's advertised models. */\n model?: string;\n } & Record;\n /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */\n ralph: {\n /** The immutable completion objective for every fresh Ralph round. */\n objective: string;\n /** Optional positive safe-integer round cap, bounded by the deployment ceiling. */\n maxRounds?: number;\n } & Record;\n /** Read a UTF-8 text file and return line-numbered content. */\n read: {\n /** Path to read, resolved by the filesystem backend. */\n file_path: string;\n /** 1-based first line to return. Defaults to 1. */\n offset?: number;\n /** Maximum number of lines to return. Defaults to 2000. */\n limit?: number;\n } & Record;\n /** Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered. */\n send_message: {\n /** The subagent id returned when the background subagent was started. */\n subagent_id: string;\n /** The message to deliver to the subagent. */\n message: string;\n } & Record;\n /** Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill. */\n skill: {\n /** The exact skill name from the available skills list. */\n name: string;\n } & Record;\n /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */\n subagent: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */\n prompt: string;\n /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */\n provider?: string;\n /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */\n model?: string;\n /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */\n reasoning_effort?: string;\n /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */\n run_in_background?: boolean;\n } & Record;\n /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result. */\n subagent_fork: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new. */\n prompt: string;\n } & Record;\n /** Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished). */\n todo_write: {\n /** The COMPLETE task list, replacing any previous list. */\n todos: ({\n /** What the task is — a short imperative line. */\n content: string;\n /** pending (not started) | in_progress (now) | completed (done). */\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n } & Record;\n /** Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn. The workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result. Script-body hooks: - `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly. - `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages. - `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`. - `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim. Misused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`. Constraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes. */\n workflow: {\n /** The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `). */\n script: string;\n /** The workflow identity block (plain JSON — never code). */\n meta: {\n /** Short kebab-case workflow name. */\n name: string;\n /** One-line description of what the workflow does. */\n description: string;\n /** Optional guidance on when this workflow applies. */\n whenToUse?: string;\n /** Optional phase declarations matched by phase() calls. */\n phases?: ({\n /** The phase title phase() calls match by exact string. */\n title: string;\n /** Optional one-line description of the phase. */\n detail?: string;\n /** Optional provider override this phase is expected to use. */\n provider?: string;\n /** Optional model override this phase is expected to use. */\n model?: string;\n } & Record)[];\n } & Record;\n /** Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}). */\n args?: Record;\n } & Record;\n /** Create or fully replace a UTF-8 text file. */\n write: {\n /** Path to write, resolved by the filesystem backend. */\n file_path: string;\n /** Full UTF-8 text content to write. */\n content: string;\n } & Record;\n}\n\ninterface ToolOutputMap {\n bash: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"foreground\";\n exitCode: number | null;\n signal: string | null;\n timedOut: boolean;\n aborted: boolean;\n timeoutMs: number;\n stdout: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n stderr: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n sandbox?: {\n mode: string;\n denied: boolean;\n enforcement?: string;\n runnerFailed?: boolean;\n };\n };\n cordis_define: {\n pluginId: string;\n packageId: string;\n name: string;\n purpose: string;\n hasHostHalf: boolean;\n hasClientHalf: boolean;\n };\n cordis_inspect_list: JsonValue;\n cordis_inspect_query: JsonValue;\n cordis_inspect_self: JsonValue;\n cordis_run: JsonValue;\n cordis_stop: {\n pluginId: string;\n };\n cordis_undefine: {\n pluginId: string;\n wasRunning: boolean;\n };\n edit: {\n path: string;\n before: string;\n after: string;\n };\n interrupt_agent: {\n accepted: boolean;\n };\n job_kill: {\n outcome: \"cancellation-requested\" | \"already-finished\";\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n job_list: ({\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n })[];\n job_output: {\n text: string;\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n list_subagent_models: string;\n ralph: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n read: {\n path: string;\n offset: number;\n lines: {\n number: number;\n text: string;\n }[];\n totalLines: number;\n };\n send_message: {\n messageId: string;\n };\n skill: {\n name: string;\n provider: string;\n resourceBase?: {\n kind: \"directory\";\n path: string;\n } | {\n kind: \"url\";\n url: string;\n } | {\n kind: \"opaque\";\n description: string;\n };\n content: string;\n };\n subagent: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n subagent_fork: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n todo_write: {\n todos: ({\n content: string;\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n counts: {\n pending: number;\n inProgress: number;\n completed: number;\n };\n };\n workflow: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n write: {\n path: string;\n operation: \"create\" | \"update\";\n before: string | null;\n after: string;\n };\n}\n\ntype ToolName = keyof ToolOutputMap\n\ndeclare class ToolCallError extends Error {\n readonly name: \"ToolCallError\";\n readonly toolName: ToolName;\n}\n\ndeclare const tools: {\n [K in ToolName]: (args: ToolArgsMap[K]) => Promise;\n}\n```","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"cordis_define","description":"Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs.","parameters":{"type":"object","properties":{"plugin":{"oneOf":[{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"new"},"idPrefix":{"type":"string","description":"Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix."}},"required":["kind","idPrefix"]},{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"existing"},"pluginId":{"type":"string","description":"Exact ID of an existing Plugin; the new Package is appended to that instance."}},"required":["kind","pluginId"]}]},"name":{"type":"string","description":"Short, readable Package name."},"purpose":{"type":"string","description":"One-sentence, user-facing description of the Package purpose."},"code":{"type":"object","additionalProperties":false,"properties":{"host":{"type":"string","description":"Plain JavaScript function body that returns the Host-half Cordis Plugin."},"client":{"type":"string","description":"Plain JavaScript function body that returns the browser Client-half Cordis Plugin."}}}},"required":["plugin","name","purpose","code"]}},{"name":"cordis_inspect_list","description":"List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call.","parameters":{"type":"object","properties":{}}},{"name":"cordis_inspect_query","description":"Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props.","parameters":{"type":"object","properties":{"platform":{"type":"string","description":"Runtime platform that owns the Provider.","enum":["host","client"]},"provider":{"type":"string","description":"Exact Provider ID returned by cordis_inspect_list."},"method":{"type":"string","description":"Exact method name declared by the Provider manifest."},"input":{"description":"Optional query input; it must satisfy the method input schema."}},"required":["platform","provider","method"]}},{"name":"cordis_inspect_self","description":"Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin."},"packageId":{"type":"string","description":"Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned."}}}},{"name":"cordis_run","description":"Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define."},"packageId":{"type":"string","description":"Exact immutable Package ID to activate under that Plugin."},"mode":{"type":"string","description":"Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package.","enum":["run","update"]}},"required":["pluginId","packageId","mode"]}},{"name":"cordis_stop","description":"Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to stop."}},"required":["pluginId"]}},{"name":"cordis_undefine","description":"Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to remove permanently."}},"required":["pluginId"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"list_subagent_models","description":"Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.","parameters":{"type":"object","properties":{"provider":{"type":"string","description":"Registered LLM provider id. Omit to list providers."},"model":{"type":"string","description":"Exact model id to inspect. Requires provider; omit to list that provider's advertised models."}}}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"run_code","description":"Execute a TypeScript program against the available tools. Takes two required arguments: `code`, the BODY of an async function (erasable syntax only; top-level `await` and `return` work), and `description`, a short summary of what the program does. Call tools as `await tools.name(args)` per the declarations in the system prompt. Only what you print or return is program output — curate it. Image-bearing subtool results are attached after the run.","parameters":{"type":"object","properties":{"code":{"type":"string","description":"The program: the body of an async TypeScript function."},"description":{"type":"string","description":"Clear, concise description of what this program does in active voice, 5-10 words (shown in the UI). Examples: \"Count TODO markers across packages\"; \"Read failing test and its fixture\"; \"Rename config key in every cordis.yml\"."}},"required":["code","description"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"provider":{"type":"string","description":"LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."},"model":{"type":"string","description":"Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."},"reasoning_effort":{"type":"string","description":"Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"text-delta","index":0,"text":"DIRECT_CHILD_OK"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"DIRECT_CHILD_OK"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"usage","usage":{"inputTokens":3,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} -{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"text","text":"DIRECT_CHILD_OK"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"02dd8a61-a39a-46d0-8f6f-457533271cae"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[10,11,12,13,14],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"text","text":"DIRECT_CHILD_OK"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"3f5bbe50-ec28-482a-ad44-729f575e12e6"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[10,11,12,13,14],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":1}} {"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} diff --git a/examples/headless-agent/tests/snapshots/advanced-toolchain/session.2.jsonl b/examples/headless-agent/tests/snapshots/advanced-toolchain/session.2.jsonl index 1f18e866a9..becebf92ee 100644 --- a/examples/headless-agent/tests/snapshots/advanced-toolchain/session.2.jsonl +++ b/examples/headless-agent/tests/snapshots/advanced-toolchain/session.2.jsonl @@ -1,19 +1,19 @@ {"type":"session","version":0,"id":"33333333-3333-4333-8333-333333333333","createdAt":1783950002000,"cwd":"{{cwd}}","parentSession":"11111111-1111-4111-8111-111111111111","origin":"subagent","delegationDepth":1} -{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly WORKFLOW_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"8b3cd23c-82f1-4903-8a3f-b9082059b40c"}]}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly WORKFLOW_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"743dc5b8-8c32-49ae-b75d-71cc45562c1a"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn"}} {"type":"step/start","data":{"turn":1,"step":1}} -{"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly WORKFLOW_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"8b3cd23c-82f1-4903-8a3f-b9082059b40c"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"823e5037-9e96-4ef5-8c5b-cbe73b993ee2"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly WORKFLOW_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"743dc5b8-8c32-49ae-b75d-71cc45562c1a"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"e7caca87-00de-4018-a7cc-627620018806"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly WORKFLOW_CHILD_OK and","messageSeqs":[5],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model. Your working directory is {{cwd}}.\n\nVerify your work by running the code or tests. Keep answers brief and factual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\n# Dynamic Cordis Plugins\n\nDynamic Cordis plugins temporarily extend the current DSH process. A Plugin uses apply(ctx) to consume Services, listen to Events, provide Services, register model Tools, or register browser UI in Slots.\n\n- Plugin and Package definitions exist only in the current process. define itself does not modify repository source, configuration, or disk, and definitions do not survive a process restart.\n- The restricted execution environment prevents accidental misuse; it is not a security boundary for malicious code. Services obtained by dynamic code connect to the real runtime.\n\n## Make the user-facing plan clear first\n\n- Dynamic Cordis Plugins are one available implementation mechanism, not the default for every request. Consider whether one could help only when the user intends to design or create something, or when a temporary interface could materially aid the current work. The presence of these instructions or Tools, and discussion of Cordis itself, do not make a request a dynamic-Plugin task.\n- When Cordis is a plausible fit, infer the intended work target and lifetime from the request and conversation. Use it only when the outcome belongs to the current running harness and should be delivered as a temporary runtime extension. If that distinction is materially ambiguous, ask at most one concise question about the intended result or lifetime. Otherwise proceed with the matching workflow; do not require the user to know or choose Cordis as an implementation mechanism.\n- Once a dynamic Plugin is appropriate, decide whether the task creates a new Plugin or modifies the Plugin named by the user with @pluginId. Proceed directly when the goal is clear; do not ask for repeated confirmation.\n- Choose Host, Client, or both from the requested outcome. Do not propose a Client/browser UI when the task does not need visible page behavior, and do not avoid Client when the requested outcome is visual, interactive, or depends on page state. Host versus Client is an implementation choice; do not make the user choose it.\n- When a design direction or a potentially useful interface would materially affect the result, ask at most one concise outcome or creative-preference question and offer a few candidate directions. Otherwise proceed directly; do not conduct a multi-round interview or a complex questionnaire.\n- cordis_define only defines and presents code; it does not run it. After definition, explain the pluginId and packageId returned by the Host and whether the next step is a run or update.\n- cordis_run may require user approval. When it returns awaiting-approval, explain that the user must allow or reject it in the UI. Do not wait, retry, or claim that it is running.\n- When it returns starting, explain that the request has entered the asynchronous flow and the Client is still activating. starting does not mean success. Wait for the system to report the final result through steering context.\n- Do not request approval again after the user rejects it. After a technical failure, fix the same Plugin from its diagnostics; do not silently create a replacement Plugin.\n\n## Recommended workflow and Tools\n\nBefore creating, modifying, or repairing a Plugin, load the cordis-plugin-development Skill. The Skill provides requirement navigation, capability composition, complete examples, and troubleshooting. Treat Inspect Provider results as the source of truth for exact APIs.\n\n1. cordis_inspect_list: discover the current Host and Client Providers and their read-only query methods.\n2. cordis_inspect_query: use the returned platform, provider, method, and schema to query exact Service, Event, Builtin, Slot, Theme token, or Tool information.\n3. cordis_inspect_self: inspect the current Session's Plugins, Packages, version pointers, source, and diagnostics. Source is returned only when both pluginId and packageId are specified.\n4. cordis_define: create the first Package for a new Plugin or append an immutable Package to an existing Plugin. It defines code but does not run it.\n5. cordis_run: activate an exact Package. Use run for the first activation, restarting current, or rollback; use update to switch versions.\n6. cordis_stop: remove the current Run and pending approval request while retaining definitions, grants, and version pointers.\n7. cordis_undefine: permanently stop and delete a Plugin and all of its Packages. Use it only after confirming that the user no longer needs them.\n\n- Inspect and Catalog data only confirm capabilities, names, signatures, types, and registration protocols before code is written; they do not replace business APIs.\n- Query Service.listService and Event.listEvents without input to choose from their compact signature directories, then query the exact service or event before using it. Exact queries return the structured contract and only its referenced types.\n- At runtime, a Plugin must call real Services or listen to real Events. Do not cache, display, or depend on Inspect results as business data.\n\n## Identity, versions, and approval\n\n- pluginId identifies a Plugin that can be modified over time. For a new Plugin, submit only a semantic idPrefix of 3–6 lowercase English letters; the Host allocates the final ID.\n- packageId identifies one immutable Host/Client source version under a Plugin. To change code, define a new Package; never overwrite an old version.\n- pluginRunId identifies one activation attempt and connects its approval, Host/Client loading, private RPC, Run card, and errors.\n- currentPackageId is the most recent fully successful Package. Stopping, starting an update, or failing an update does not clear it.\n- nextPackageId is the target awaiting approval, being attempted, awaiting Client activation, or most recently failed.\n- A single check mark authorizes only the current Package; double check marks authorize future versions of the same Plugin. A grant remains in effect after a technical failure.\n- An update stops the old Run before starting the target Package. Failure does not automatically restart the old version; retry next with update or roll back to current with run.\n\nWhen the user enters @pluginId, the system injects identity, the default base Package, version pointers, and runtime status, but not source code:\n\n1. Call cordis_inspect_self(pluginId, packageId) to read the target source.\n2. Use cordis_define in existing mode to append a Package to the same Plugin.\n3. Call cordis_run in run or update mode according to the version relationship.\n\nNever silently create another Plugin for @pluginId. If the reference is unavailable because it was removed, belongs to another Session, or was lost on process restart, tell the user directly.\n\n## High-frequency errors that must be avoided\n\n### Services: ctx.get and inject\n\n- Read an optional Service with ctx.get('serviceName') by default and handle undefined.\n- Declare inject: ['serviceName'] on the returned Plugin object only when the Service is a hard dependency and the Plugin must enter waiting until Cordis reactivates it after the Service appears.\n- Read ctx.serviceName only after declaring that Service in inject. Never access an undeclared Service as a ctx property.\n\n```js\nreturn {\n inject: ['requiredService'],\n apply(ctx) {\n ctx.requiredService.someMethod()\n const optionalService = ctx.get('optionalService')\n if (optionalService !== undefined) optionalService.someMethod()\n },\n}\n```\n\n### Code: use plain JavaScript only\n\n- Host and Client code is not transformed by TypeScript, JSX, or a bundler.\n- Do not use TypeScript types, as, decorators, import, require, or JSX.\n- Client React code must use React.createElement(...); never write .\n- Do not assume that process, Buffer, window, document, fetch, native timers, or any other global is available. Query the corresponding platform's Builtins and Services first.\n\n### Data: do not serialize live data\n\n- Services, Events, Slots, Sessions, and their derived Cordis/DSH objects are internal live data, not ordinary JSON that can be dumped.\n- Do not apply JSON.stringify, structuredClone, recursive enumeration, full copying, or whole-object display to live data.\n- Read only the leaf fields required by the task, then construct the smallest owned data object without Host references.\n\n### Lifecycle: every side effect must be reversible\n\n- Services, Events, Tools, handlers, timers, Slots, styles, and theme overrides must all belong to the current Fiber.\n- Use ctx.effect(), ctx.on(), or official APIs that return a disposer so stop, update, or undefine removes every side effect.\n- The cordis-plugin-development Skill contains complete timer, Waterfall, Slot, theme, Tool, RPC, and React examples and troubleshooting guidance.\n\n## Host and Client\n\n- Host runs in the DSH Node.js process and is appropriate for files, networking, commands, Agent/Session access, Host Events, Services, model Tools, and JSON methods callable by the Client.\n- Client runs in the browser page and is appropriate for themes, layout, current page state, Tool cards, and Slot UI.\n- Host and Client communicate through Package-private JSON methods: Host uses harness.handle(method, handler), and Client uses host.call(method, args). The direction is Client→Host, and only lossless JSON may cross it.\n- Client UI must be registered in a queried Slot; apply() cannot directly return a React Element. Query Slots.listSubTree without root to choose from the compact purpose/topology tree, then query the exact root for its full registration contract and props before writing code.\n- See the Skill and Inspect Providers for Run-specific panels and exact Slot registration patterns.\n\n## Asynchronous results and recovery\n\n- Do not wait inside a Tool for approval or browser work that can happen only after the current turn ends.\n- Asynchronous success, rejection, and runtime errors update Run state and notify you through steering context.\n- After a technical failure, use cordis_inspect_self to read the exact Package source and its message/stack. Define a corrected Package under the same Plugin and retry autonomously.\n- Use the cordis-plugin-development Skill for other failure causes, repair procedures, and complete extension patterns.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.\n\n## Writing code for run_code\n\n`run_code` takes two required arguments: `code` — the body of an async TypeScript function (erasable syntax only — no `enum` or namespaces; type annotations are advisory, the code runs type-stripped) — and `description`, a short summary of what the program does. Inside the program:\n\n- Call tools as `await tools.name(args)` — quoted access for exotic names: `tools[\"my-tool\"](args)`. Every call resolves to the tool's typed canonical JSON value. Tool arguments must be lossless JSON.\n- A FAILED tool call rejects with `ToolCallError`, whose `toolName` identifies the failed tool and whose `message` is human-readable — `try/catch` it to handle and continue.\n- Independent read-only calls MAY overlap under `Promise.all` (safe calls run concurrently; mutating calls run alone, in submission order). Sequence dependent work with `await`.\n- Emit results with `return` and/or `console.log(...)`. ONLY what you print or return comes back to you — intermediate tool results never enter the conversation, so extract just what you need.\n\nThe available tools:\n\n```ts\ntype JsonValue = null | boolean | number | string | JsonValue[] | { [key: string]: JsonValue }\n\ninterface ToolArgsMap {\n /** Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`. */\n bash: {\n /** The bash command to execute. */\n command: string;\n /** Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\". */\n description: string;\n /** Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry. */\n timeoutMs?: number;\n /** Working directory for this command. Defaults to the session workspace; a relative path is resolved against it. */\n workdir?: string;\n /** Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies. */\n run_in_background?: boolean;\n } & Record;\n /** Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs. */\n cordis_define: {\n plugin: {\n kind: \"new\";\n /** Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix. */\n idPrefix: string;\n } | {\n kind: \"existing\";\n /** Exact ID of an existing Plugin; the new Package is appended to that instance. */\n pluginId: string;\n };\n /** Short, readable Package name. */\n name: string;\n /** One-sentence, user-facing description of the Package purpose. */\n purpose: string;\n code: {\n /** Plain JavaScript function body that returns the Host-half Cordis Plugin. */\n host?: string;\n /** Plain JavaScript function body that returns the browser Client-half Cordis Plugin. */\n client?: string;\n };\n } & Record;\n /** List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call. */\n cordis_inspect_list: Record;\n /** Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props. */\n cordis_inspect_query: {\n /** Runtime platform that owns the Provider. */\n platform: \"host\" | \"client\";\n /** Exact Provider ID returned by cordis_inspect_list. */\n provider: string;\n /** Exact method name declared by the Provider manifest. */\n method: string;\n /** Optional query input; it must satisfy the method input schema. */\n input?: JsonValue;\n } & Record;\n /** Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers. */\n cordis_inspect_self: {\n /** Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin. */\n pluginId?: string;\n /** Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned. */\n packageId?: string;\n } & Record;\n /** Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it. */\n cordis_run: {\n /** Stable Plugin ID returned by cordis_define. */\n pluginId: string;\n /** Exact immutable Package ID to activate under that Plugin. */\n packageId: string;\n /** Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package. */\n mode: \"run\" | \"update\";\n } & Record;\n /** Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal. */\n cordis_stop: {\n /** Stable dynamic Plugin ID to stop. */\n pluginId: string;\n } & Record;\n /** Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead. */\n cordis_undefine: {\n /** Stable dynamic Plugin ID to remove permanently. */\n pluginId: string;\n } & Record;\n /** Edit an existing UTF-8 text file by replacing literal text. */\n edit: {\n /** Path to edit, resolved by the filesystem backend. */\n file_path: string;\n /** Literal text to replace. Must match exactly. */\n old_string: string;\n /** Literal replacement text. Use an empty string to delete the match. */\n new_string: string;\n /** Replace all matches. Defaults to false; when false, old_string must appear exactly once. */\n replace_all?: boolean;\n } & Record;\n /** Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op. */\n interrupt_agent: {\n /** The agent id of the running agent to interrupt. */\n agent_id: string;\n } & Record;\n /** Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops. */\n job_kill: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Optional short reason, recorded in the log and forwarded to the job. */\n reason?: string;\n } & Record;\n /** List your background jobs (running and finished) with their ids, kinds, and statuses. */\n job_list: Record;\n /** Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap. */\n job_output: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive. */\n wait?: boolean;\n /** Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum. */\n timeout_ms?: number;\n } & Record;\n /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */\n ralph: {\n /** The immutable completion objective for every fresh Ralph round. */\n objective: string;\n /** Optional positive safe-integer round cap, bounded by the deployment ceiling. */\n maxRounds?: number;\n } & Record;\n /** Read a UTF-8 text file and return line-numbered content. */\n read: {\n /** Path to read, resolved by the filesystem backend. */\n file_path: string;\n /** 1-based first line to return. Defaults to 1. */\n offset?: number;\n /** Maximum number of lines to return. Defaults to 2000. */\n limit?: number;\n } & Record;\n /** Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered. */\n send_message: {\n /** The subagent id returned when the background subagent was started. */\n subagent_id: string;\n /** The message to deliver to the subagent. */\n message: string;\n } & Record;\n /** Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill. */\n skill: {\n /** The exact skill name from the available skills list. */\n name: string;\n } & Record;\n /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. */\n subagent: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */\n prompt: string;\n /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */\n run_in_background?: boolean;\n } & Record;\n /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result. */\n subagent_fork: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new. */\n prompt: string;\n } & Record;\n /** Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished). */\n todo_write: {\n /** The COMPLETE task list, replacing any previous list. */\n todos: ({\n /** What the task is — a short imperative line. */\n content: string;\n /** pending (not started) | in_progress (now) | completed (done). */\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n } & Record;\n /** Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn. The workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result. Script-body hooks: - `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly. - `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages. - `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`. - `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim. Misused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`. Constraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes. */\n workflow: {\n /** The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `). */\n script: string;\n /** The workflow identity block (plain JSON — never code). */\n meta: {\n /** Short kebab-case workflow name. */\n name: string;\n /** One-line description of what the workflow does. */\n description: string;\n /** Optional guidance on when this workflow applies. */\n whenToUse?: string;\n /** Optional phase declarations matched by phase() calls. */\n phases?: ({\n /** The phase title phase() calls match by exact string. */\n title: string;\n /** Optional one-line description of the phase. */\n detail?: string;\n /** Optional provider override this phase is expected to use. */\n provider?: string;\n /** Optional model override this phase is expected to use. */\n model?: string;\n } & Record)[];\n } & Record;\n /** Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}). */\n args?: Record;\n } & Record;\n /** Create or fully replace a UTF-8 text file. */\n write: {\n /** Path to write, resolved by the filesystem backend. */\n file_path: string;\n /** Full UTF-8 text content to write. */\n content: string;\n } & Record;\n}\n\ninterface ToolOutputMap {\n bash: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"foreground\";\n exitCode: number | null;\n signal: string | null;\n timedOut: boolean;\n aborted: boolean;\n timeoutMs: number;\n stdout: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n stderr: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n sandbox?: {\n mode: string;\n denied: boolean;\n enforcement?: string;\n runnerFailed?: boolean;\n };\n };\n cordis_define: {\n pluginId: string;\n packageId: string;\n name: string;\n purpose: string;\n hasHostHalf: boolean;\n hasClientHalf: boolean;\n };\n cordis_inspect_list: JsonValue;\n cordis_inspect_query: JsonValue;\n cordis_inspect_self: JsonValue;\n cordis_run: JsonValue;\n cordis_stop: {\n pluginId: string;\n };\n cordis_undefine: {\n pluginId: string;\n wasRunning: boolean;\n };\n edit: {\n path: string;\n before: string;\n after: string;\n };\n interrupt_agent: {\n accepted: boolean;\n };\n job_kill: {\n outcome: \"cancellation-requested\" | \"already-finished\";\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n job_list: ({\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n })[];\n job_output: {\n text: string;\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n ralph: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n read: {\n path: string;\n offset: number;\n lines: {\n number: number;\n text: string;\n }[];\n totalLines: number;\n };\n send_message: {\n messageId: string;\n };\n skill: {\n name: string;\n provider: string;\n resourceBase?: {\n kind: \"directory\";\n path: string;\n } | {\n kind: \"url\";\n url: string;\n } | {\n kind: \"opaque\";\n description: string;\n };\n content: string;\n };\n subagent: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n subagent_fork: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n todo_write: {\n todos: ({\n content: string;\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n counts: {\n pending: number;\n inProgress: number;\n completed: number;\n };\n };\n workflow: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n write: {\n path: string;\n operation: \"create\" | \"update\";\n before: string | null;\n after: string;\n };\n}\n\ntype ToolName = keyof ToolOutputMap\n\ndeclare class ToolCallError extends Error {\n readonly name: \"ToolCallError\";\n readonly toolName: ToolName;\n}\n\ndeclare const tools: {\n [K in ToolName]: (args: ToolArgsMap[K]) => Promise;\n}\n```","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"cordis_define","description":"Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs.","parameters":{"type":"object","properties":{"plugin":{"oneOf":[{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"new"},"idPrefix":{"type":"string","description":"Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix."}},"required":["kind","idPrefix"]},{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"existing"},"pluginId":{"type":"string","description":"Exact ID of an existing Plugin; the new Package is appended to that instance."}},"required":["kind","pluginId"]}]},"name":{"type":"string","description":"Short, readable Package name."},"purpose":{"type":"string","description":"One-sentence, user-facing description of the Package purpose."},"code":{"type":"object","additionalProperties":false,"properties":{"host":{"type":"string","description":"Plain JavaScript function body that returns the Host-half Cordis Plugin."},"client":{"type":"string","description":"Plain JavaScript function body that returns the browser Client-half Cordis Plugin."}}}},"required":["plugin","name","purpose","code"]}},{"name":"cordis_inspect_list","description":"List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call.","parameters":{"type":"object","properties":{}}},{"name":"cordis_inspect_query","description":"Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props.","parameters":{"type":"object","properties":{"platform":{"type":"string","description":"Runtime platform that owns the Provider.","enum":["host","client"]},"provider":{"type":"string","description":"Exact Provider ID returned by cordis_inspect_list."},"method":{"type":"string","description":"Exact method name declared by the Provider manifest."},"input":{"description":"Optional query input; it must satisfy the method input schema."}},"required":["platform","provider","method"]}},{"name":"cordis_inspect_self","description":"Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin."},"packageId":{"type":"string","description":"Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned."}}}},{"name":"cordis_run","description":"Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define."},"packageId":{"type":"string","description":"Exact immutable Package ID to activate under that Plugin."},"mode":{"type":"string","description":"Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package.","enum":["run","update"]}},"required":["pluginId","packageId","mode"]}},{"name":"cordis_stop","description":"Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to stop."}},"required":["pluginId"]}},{"name":"cordis_undefine","description":"Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to remove permanently."}},"required":["pluginId"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"run_code","description":"Execute a TypeScript program against the available tools. Takes two required arguments: `code`, the BODY of an async function (erasable syntax only; top-level `await` and `return` work), and `description`, a short summary of what the program does. Call tools as `await tools.name(args)` per the declarations in the system prompt. Only what you print or return comes back — curate it.","parameters":{"type":"object","properties":{"code":{"type":"string","description":"The program: the body of an async TypeScript function."},"description":{"type":"string","description":"Clear, concise description of what this program does in active voice, 5-10 words (shown in the UI). Examples: \"Count TODO markers across packages\"; \"Read failing test and its fixture\"; \"Rename config key in every cordis.yml\"."}},"required":["code","description"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model. Your working directory is {{cwd}}.\n\nVerify your work by running the code or tests. Keep answers brief and factual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\n# Dynamic Cordis Plugins\n\nDynamic Cordis plugins temporarily extend the current DSH process. A Plugin uses apply(ctx) to consume Services, listen to Events, provide Services, register model Tools, or register browser UI in Slots.\n\n- Plugin and Package definitions exist only in the current process. define itself does not modify repository source, configuration, or disk, and definitions do not survive a process restart.\n- The restricted execution environment prevents accidental misuse; it is not a security boundary for malicious code. Services obtained by dynamic code connect to the real runtime.\n\n## Make the user-facing plan clear first\n\n- Dynamic Cordis Plugins are one available implementation mechanism, not the default for every request. Consider whether one could help only when the user intends to design or create something, or when a temporary interface could materially aid the current work. The presence of these instructions or Tools, and discussion of Cordis itself, do not make a request a dynamic-Plugin task.\n- When Cordis is a plausible fit, infer the intended work target and lifetime from the request and conversation. Use it only when the outcome belongs to the current running harness and should be delivered as a temporary runtime extension. If that distinction is materially ambiguous, ask at most one concise question about the intended result or lifetime. Otherwise proceed with the matching workflow; do not require the user to know or choose Cordis as an implementation mechanism.\n- Once a dynamic Plugin is appropriate, decide whether the task creates a new Plugin or modifies the Plugin named by the user with @pluginId. Proceed directly when the goal is clear; do not ask for repeated confirmation.\n- Choose Host, Client, or both from the requested outcome. Do not propose a Client/browser UI when the task does not need visible page behavior, and do not avoid Client when the requested outcome is visual, interactive, or depends on page state. Host versus Client is an implementation choice; do not make the user choose it.\n- When a design direction or a potentially useful interface would materially affect the result, ask at most one concise outcome or creative-preference question and offer a few candidate directions. Otherwise proceed directly; do not conduct a multi-round interview or a complex questionnaire.\n- cordis_define only defines and presents code; it does not run it. After definition, explain the pluginId and packageId returned by the Host and whether the next step is a run or update.\n- cordis_run may require user approval. When it returns awaiting-approval, explain that the user must allow or reject it in the UI. Do not wait, retry, or claim that it is running.\n- When it returns starting, explain that the request has entered the asynchronous flow and the Client is still activating. starting does not mean success. Wait for the system to report the final result through steering context.\n- Do not request approval again after the user rejects it. After a technical failure, fix the same Plugin from its diagnostics; do not silently create a replacement Plugin.\n\n## Recommended workflow and Tools\n\nBefore creating, modifying, or repairing a Plugin, load the cordis-plugin-development Skill. The Skill provides requirement navigation, capability composition, complete examples, and troubleshooting. Treat Inspect Provider results as the source of truth for exact APIs.\n\n1. cordis_inspect_list: discover the current Host and Client Providers and their read-only query methods.\n2. cordis_inspect_query: use the returned platform, provider, method, and schema to query exact Service, Event, Builtin, Slot, Theme token, or Tool information.\n3. cordis_inspect_self: inspect the current Session's Plugins, Packages, version pointers, source, and diagnostics. Source is returned only when both pluginId and packageId are specified.\n4. cordis_define: create the first Package for a new Plugin or append an immutable Package to an existing Plugin. It defines code but does not run it.\n5. cordis_run: activate an exact Package. Use run for the first activation, restarting current, or rollback; use update to switch versions.\n6. cordis_stop: remove the current Run and pending approval request while retaining definitions, grants, and version pointers.\n7. cordis_undefine: permanently stop and delete a Plugin and all of its Packages. Use it only after confirming that the user no longer needs them.\n\n- Inspect and Catalog data only confirm capabilities, names, signatures, types, and registration protocols before code is written; they do not replace business APIs.\n- Query Service.listService and Event.listEvents without input to choose from their compact signature directories, then query the exact service or event before using it. Exact queries return the structured contract and only its referenced types.\n- At runtime, a Plugin must call real Services or listen to real Events. Do not cache, display, or depend on Inspect results as business data.\n\n## Identity, versions, and approval\n\n- pluginId identifies a Plugin that can be modified over time. For a new Plugin, submit only a semantic idPrefix of 3–6 lowercase English letters; the Host allocates the final ID.\n- packageId identifies one immutable Host/Client source version under a Plugin. To change code, define a new Package; never overwrite an old version.\n- pluginRunId identifies one activation attempt and connects its approval, Host/Client loading, private RPC, Run card, and errors.\n- currentPackageId is the most recent fully successful Package. Stopping, starting an update, or failing an update does not clear it.\n- nextPackageId is the target awaiting approval, being attempted, awaiting Client activation, or most recently failed.\n- A single check mark authorizes only the current Package; double check marks authorize future versions of the same Plugin. A grant remains in effect after a technical failure.\n- An update stops the old Run before starting the target Package. Failure does not automatically restart the old version; retry next with update or roll back to current with run.\n\nWhen the user enters @pluginId, the system injects identity, the default base Package, version pointers, and runtime status, but not source code:\n\n1. Call cordis_inspect_self(pluginId, packageId) to read the target source.\n2. Use cordis_define in existing mode to append a Package to the same Plugin.\n3. Call cordis_run in run or update mode according to the version relationship.\n\nNever silently create another Plugin for @pluginId. If the reference is unavailable because it was removed, belongs to another Session, or was lost on process restart, tell the user directly.\n\n## High-frequency errors that must be avoided\n\n### Services: ctx.get and inject\n\n- Read an optional Service with ctx.get('serviceName') by default and handle undefined.\n- Declare inject: ['serviceName'] on the returned Plugin object only when the Service is a hard dependency and the Plugin must enter waiting until Cordis reactivates it after the Service appears.\n- Read ctx.serviceName only after declaring that Service in inject. Never access an undeclared Service as a ctx property.\n\n```js\nreturn {\n inject: ['requiredService'],\n apply(ctx) {\n ctx.requiredService.someMethod()\n const optionalService = ctx.get('optionalService')\n if (optionalService !== undefined) optionalService.someMethod()\n },\n}\n```\n\n### Code: use plain JavaScript only\n\n- Host and Client code is not transformed by TypeScript, JSX, or a bundler.\n- Do not use TypeScript types, as, decorators, import, require, or JSX.\n- Client React code must use React.createElement(...); never write .\n- Do not assume that process, Buffer, window, document, fetch, native timers, or any other global is available. Query the corresponding platform's Builtins and Services first.\n\n### Data: do not serialize live data\n\n- Services, Events, Slots, Sessions, and their derived Cordis/DSH objects are internal live data, not ordinary JSON that can be dumped.\n- Do not apply JSON.stringify, structuredClone, recursive enumeration, full copying, or whole-object display to live data.\n- Read only the leaf fields required by the task, then construct the smallest owned data object without Host references.\n\n### Lifecycle: every side effect must be reversible\n\n- Services, Events, Tools, handlers, timers, Slots, styles, and theme overrides must all belong to the current Fiber.\n- Use ctx.effect(), ctx.on(), or official APIs that return a disposer so stop, update, or undefine removes every side effect.\n- The cordis-plugin-development Skill contains complete timer, Waterfall, Slot, theme, Tool, RPC, and React examples and troubleshooting guidance.\n\n## Host and Client\n\n- Host runs in the DSH Node.js process and is appropriate for files, networking, commands, Agent/Session access, Host Events, Services, model Tools, and JSON methods callable by the Client.\n- Client runs in the browser page and is appropriate for themes, layout, current page state, Tool cards, and Slot UI.\n- Host and Client communicate through Package-private JSON methods: Host uses harness.handle(method, handler), and Client uses host.call(method, args). The direction is Client→Host, and only lossless JSON may cross it.\n- Client UI must be registered in a queried Slot; apply() cannot directly return a React Element. Query Slots.listSubTree without root to choose from the compact purpose/topology tree, then query the exact root for its full registration contract and props before writing code.\n- See the Skill and Inspect Providers for Run-specific panels and exact Slot registration patterns.\n\n## Asynchronous results and recovery\n\n- Do not wait inside a Tool for approval or browser work that can happen only after the current turn ends.\n- Asynchronous success, rejection, and runtime errors update Run state and notify you through steering context.\n- After a technical failure, use cordis_inspect_self to read the exact Package source and its message/stack. Define a corrected Package under the same Plugin and retry autonomously.\n- Use the cordis-plugin-development Skill for other failure causes, repair procedures, and complete extension patterns.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.\n\n## Writing code for run_code\n\n`run_code` takes two required arguments: `code` — the body of an async TypeScript function (erasable syntax only — no `enum` or namespaces; type annotations are advisory, the code runs type-stripped) — and `description`, a short summary of what the program does. Inside the program:\n\n- Call tools as `await tools.name(args)` — quoted access for exotic names: `tools[\"my-tool\"](args)`. Every call resolves to the tool's typed canonical JSON value. Tool arguments must be lossless JSON.\n- A FAILED tool call rejects with `ToolCallError`, whose `toolName` identifies the failed tool and whose `message` is human-readable — `try/catch` it to handle and continue.\n- Independent read-only calls MAY overlap under `Promise.all` (safe calls run concurrently; mutating calls run alone, in submission order). Sequence dependent work with `await`.\n- Emit results with `return` and/or `console.log(...)`. Only what you print or return is program output. A successful tool result containing an image is attached after the run so you can inspect it on the next step; every other intermediate result stays out of the conversation, so extract just what you need.\n\nThe available tools:\n\n```ts\ntype JsonValue = null | boolean | number | string | JsonValue[] | { [key: string]: JsonValue }\n\ninterface ToolArgsMap {\n /** Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`. */\n bash: {\n /** The bash command to execute. */\n command: string;\n /** Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\". */\n description: string;\n /** Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry. */\n timeoutMs?: number;\n /** Working directory for this command. Defaults to the session workspace; a relative path is resolved against it. */\n workdir?: string;\n /** Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies. */\n run_in_background?: boolean;\n } & Record;\n /** Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs. */\n cordis_define: {\n plugin: {\n kind: \"new\";\n /** Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix. */\n idPrefix: string;\n } | {\n kind: \"existing\";\n /** Exact ID of an existing Plugin; the new Package is appended to that instance. */\n pluginId: string;\n };\n /** Short, readable Package name. */\n name: string;\n /** One-sentence, user-facing description of the Package purpose. */\n purpose: string;\n code: {\n /** Plain JavaScript function body that returns the Host-half Cordis Plugin. */\n host?: string;\n /** Plain JavaScript function body that returns the browser Client-half Cordis Plugin. */\n client?: string;\n };\n } & Record;\n /** List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call. */\n cordis_inspect_list: Record;\n /** Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props. */\n cordis_inspect_query: {\n /** Runtime platform that owns the Provider. */\n platform: \"host\" | \"client\";\n /** Exact Provider ID returned by cordis_inspect_list. */\n provider: string;\n /** Exact method name declared by the Provider manifest. */\n method: string;\n /** Optional query input; it must satisfy the method input schema. */\n input?: JsonValue;\n } & Record;\n /** Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers. */\n cordis_inspect_self: {\n /** Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin. */\n pluginId?: string;\n /** Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned. */\n packageId?: string;\n } & Record;\n /** Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it. */\n cordis_run: {\n /** Stable Plugin ID returned by cordis_define. */\n pluginId: string;\n /** Exact immutable Package ID to activate under that Plugin. */\n packageId: string;\n /** Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package. */\n mode: \"run\" | \"update\";\n } & Record;\n /** Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal. */\n cordis_stop: {\n /** Stable dynamic Plugin ID to stop. */\n pluginId: string;\n } & Record;\n /** Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead. */\n cordis_undefine: {\n /** Stable dynamic Plugin ID to remove permanently. */\n pluginId: string;\n } & Record;\n /** Edit an existing UTF-8 text file by replacing literal text. */\n edit: {\n /** Path to edit, resolved by the filesystem backend. */\n file_path: string;\n /** Literal text to replace. Must match exactly. */\n old_string: string;\n /** Literal replacement text. Use an empty string to delete the match. */\n new_string: string;\n /** Replace all matches. Defaults to false; when false, old_string must appear exactly once. */\n replace_all?: boolean;\n } & Record;\n /** Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op. */\n interrupt_agent: {\n /** The agent id of the running agent to interrupt. */\n agent_id: string;\n } & Record;\n /** Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops. */\n job_kill: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Optional short reason, recorded in the log and forwarded to the job. */\n reason?: string;\n } & Record;\n /** List your background jobs (running and finished) with their ids, kinds, and statuses. */\n job_list: Record;\n /** Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap. */\n job_output: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive. */\n wait?: boolean;\n /** Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum. */\n timeout_ms?: number;\n } & Record;\n /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */\n list_subagent_models: {\n /** Registered LLM provider id. Omit to list providers. */\n provider?: string;\n /** Exact model id to inspect. Requires provider; omit to list that provider's advertised models. */\n model?: string;\n } & Record;\n /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */\n ralph: {\n /** The immutable completion objective for every fresh Ralph round. */\n objective: string;\n /** Optional positive safe-integer round cap, bounded by the deployment ceiling. */\n maxRounds?: number;\n } & Record;\n /** Read a UTF-8 text file and return line-numbered content. */\n read: {\n /** Path to read, resolved by the filesystem backend. */\n file_path: string;\n /** 1-based first line to return. Defaults to 1. */\n offset?: number;\n /** Maximum number of lines to return. Defaults to 2000. */\n limit?: number;\n } & Record;\n /** Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered. */\n send_message: {\n /** The subagent id returned when the background subagent was started. */\n subagent_id: string;\n /** The message to deliver to the subagent. */\n message: string;\n } & Record;\n /** Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill. */\n skill: {\n /** The exact skill name from the available skills list. */\n name: string;\n } & Record;\n /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */\n subagent: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */\n prompt: string;\n /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */\n provider?: string;\n /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */\n model?: string;\n /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */\n reasoning_effort?: string;\n /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */\n run_in_background?: boolean;\n } & Record;\n /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result. */\n subagent_fork: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new. */\n prompt: string;\n } & Record;\n /** Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished). */\n todo_write: {\n /** The COMPLETE task list, replacing any previous list. */\n todos: ({\n /** What the task is — a short imperative line. */\n content: string;\n /** pending (not started) | in_progress (now) | completed (done). */\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n } & Record;\n /** Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn. The workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result. Script-body hooks: - `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly. - `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages. - `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`. - `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim. Misused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`. Constraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes. */\n workflow: {\n /** The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `). */\n script: string;\n /** The workflow identity block (plain JSON — never code). */\n meta: {\n /** Short kebab-case workflow name. */\n name: string;\n /** One-line description of what the workflow does. */\n description: string;\n /** Optional guidance on when this workflow applies. */\n whenToUse?: string;\n /** Optional phase declarations matched by phase() calls. */\n phases?: ({\n /** The phase title phase() calls match by exact string. */\n title: string;\n /** Optional one-line description of the phase. */\n detail?: string;\n /** Optional provider override this phase is expected to use. */\n provider?: string;\n /** Optional model override this phase is expected to use. */\n model?: string;\n } & Record)[];\n } & Record;\n /** Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}). */\n args?: Record;\n } & Record;\n /** Create or fully replace a UTF-8 text file. */\n write: {\n /** Path to write, resolved by the filesystem backend. */\n file_path: string;\n /** Full UTF-8 text content to write. */\n content: string;\n } & Record;\n}\n\ninterface ToolOutputMap {\n bash: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"foreground\";\n exitCode: number | null;\n signal: string | null;\n timedOut: boolean;\n aborted: boolean;\n timeoutMs: number;\n stdout: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n stderr: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n sandbox?: {\n mode: string;\n denied: boolean;\n enforcement?: string;\n runnerFailed?: boolean;\n };\n };\n cordis_define: {\n pluginId: string;\n packageId: string;\n name: string;\n purpose: string;\n hasHostHalf: boolean;\n hasClientHalf: boolean;\n };\n cordis_inspect_list: JsonValue;\n cordis_inspect_query: JsonValue;\n cordis_inspect_self: JsonValue;\n cordis_run: JsonValue;\n cordis_stop: {\n pluginId: string;\n };\n cordis_undefine: {\n pluginId: string;\n wasRunning: boolean;\n };\n edit: {\n path: string;\n before: string;\n after: string;\n };\n interrupt_agent: {\n accepted: boolean;\n };\n job_kill: {\n outcome: \"cancellation-requested\" | \"already-finished\";\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n job_list: ({\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n })[];\n job_output: {\n text: string;\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n list_subagent_models: string;\n ralph: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n read: {\n path: string;\n offset: number;\n lines: {\n number: number;\n text: string;\n }[];\n totalLines: number;\n };\n send_message: {\n messageId: string;\n };\n skill: {\n name: string;\n provider: string;\n resourceBase?: {\n kind: \"directory\";\n path: string;\n } | {\n kind: \"url\";\n url: string;\n } | {\n kind: \"opaque\";\n description: string;\n };\n content: string;\n };\n subagent: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n subagent_fork: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n todo_write: {\n todos: ({\n content: string;\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n counts: {\n pending: number;\n inProgress: number;\n completed: number;\n };\n };\n workflow: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n write: {\n path: string;\n operation: \"create\" | \"update\";\n before: string | null;\n after: string;\n };\n}\n\ntype ToolName = keyof ToolOutputMap\n\ndeclare class ToolCallError extends Error {\n readonly name: \"ToolCallError\";\n readonly toolName: ToolName;\n}\n\ndeclare const tools: {\n [K in ToolName]: (args: ToolArgsMap[K]) => Promise;\n}\n```","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"cordis_define","description":"Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs.","parameters":{"type":"object","properties":{"plugin":{"oneOf":[{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"new"},"idPrefix":{"type":"string","description":"Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix."}},"required":["kind","idPrefix"]},{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"existing"},"pluginId":{"type":"string","description":"Exact ID of an existing Plugin; the new Package is appended to that instance."}},"required":["kind","pluginId"]}]},"name":{"type":"string","description":"Short, readable Package name."},"purpose":{"type":"string","description":"One-sentence, user-facing description of the Package purpose."},"code":{"type":"object","additionalProperties":false,"properties":{"host":{"type":"string","description":"Plain JavaScript function body that returns the Host-half Cordis Plugin."},"client":{"type":"string","description":"Plain JavaScript function body that returns the browser Client-half Cordis Plugin."}}}},"required":["plugin","name","purpose","code"]}},{"name":"cordis_inspect_list","description":"List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call.","parameters":{"type":"object","properties":{}}},{"name":"cordis_inspect_query","description":"Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props.","parameters":{"type":"object","properties":{"platform":{"type":"string","description":"Runtime platform that owns the Provider.","enum":["host","client"]},"provider":{"type":"string","description":"Exact Provider ID returned by cordis_inspect_list."},"method":{"type":"string","description":"Exact method name declared by the Provider manifest."},"input":{"description":"Optional query input; it must satisfy the method input schema."}},"required":["platform","provider","method"]}},{"name":"cordis_inspect_self","description":"Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin."},"packageId":{"type":"string","description":"Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned."}}}},{"name":"cordis_run","description":"Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define."},"packageId":{"type":"string","description":"Exact immutable Package ID to activate under that Plugin."},"mode":{"type":"string","description":"Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package.","enum":["run","update"]}},"required":["pluginId","packageId","mode"]}},{"name":"cordis_stop","description":"Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to stop."}},"required":["pluginId"]}},{"name":"cordis_undefine","description":"Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to remove permanently."}},"required":["pluginId"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"list_subagent_models","description":"Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.","parameters":{"type":"object","properties":{"provider":{"type":"string","description":"Registered LLM provider id. Omit to list providers."},"model":{"type":"string","description":"Exact model id to inspect. Requires provider; omit to list that provider's advertised models."}}}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"run_code","description":"Execute a TypeScript program against the available tools. Takes two required arguments: `code`, the BODY of an async function (erasable syntax only; top-level `await` and `return` work), and `description`, a short summary of what the program does. Call tools as `await tools.name(args)` per the declarations in the system prompt. Only what you print or return is program output — curate it. Image-bearing subtool results are attached after the run.","parameters":{"type":"object","properties":{"code":{"type":"string","description":"The program: the body of an async TypeScript function."},"description":{"type":"string","description":"Clear, concise description of what this program does in active voice, 5-10 words (shown in the UI). Examples: \"Count TODO markers across packages\"; \"Read failing test and its fixture\"; \"Rename config key in every cordis.yml\"."}},"required":["code","description"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"provider":{"type":"string","description":"LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."},"model":{"type":"string","description":"Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."},"reasoning_effort":{"type":"string","description":"Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"text-delta","index":0,"text":"WORKFLOW_CHILD_OK"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"WORKFLOW_CHILD_OK"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"usage","usage":{"inputTokens":3,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} -{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"text","text":"WORKFLOW_CHILD_OK"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"cd78f077-1fad-4cdc-ab56-09d39d9095cd"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[10,11,12,13,14],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"text","text":"WORKFLOW_CHILD_OK"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"c07ab6db-7108-46b4-ad18-3f81e9609546"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[10,11,12,13,14],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":1}} {"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} diff --git a/examples/headless-agent/tests/snapshots/advanced-toolchain/session.jsonl b/examples/headless-agent/tests/snapshots/advanced-toolchain/session.jsonl index c49d9b2745..4bed94ecda 100644 --- a/examples/headless-agent/tests/snapshots/advanced-toolchain/session.jsonl +++ b/examples/headless-agent/tests/snapshots/advanced-toolchain/session.jsonl @@ -1,20 +1,20 @@ {"type":"session","version":0,"id":"11111111-1111-4111-8111-111111111111","createdAt":1783950000000,"cwd":"{{cwd}}","delegationDepth":0} -{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Run this advanced flow exactly once: define a host-only dynamic Cordis Package named Snapshot Marker; run and inspect snap-1/pkg-1 through run_code; delegate once to a direct spawn child; run one workflow that delegates to another spawn child; remove snap-1; then reply with exactly ADVANCED_HEADLESS_OK."}],"source":{"kind":"user"},"role":"user","id":"8a0ac233-283c-4eb4-8bbd-5c50b7e99afe"}]}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Run this advanced flow exactly once: define a host-only dynamic Cordis Package named Snapshot Marker; run and inspect snap-1/pkg-1 through run_code; delegate once to a direct spawn child; run one workflow that delegates to another spawn child; remove snap-1; then reply with exactly ADVANCED_HEADLESS_OK."}],"source":{"kind":"user"},"role":"user","id":"e203969f-330f-4886-8d34-82d0001db89b"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} {"type":"step/start","data":{"turn":1,"step":1}} -{"type":"user/message","data":{"content":[{"type":"text","text":"Run this advanced flow exactly once: define a host-only dynamic Cordis Package named Snapshot Marker; run and inspect snap-1/pkg-1 through run_code; delegate once to a direct spawn child; run one workflow that delegates to another spawn child; remove snap-1; then reply with exactly ADVANCED_HEADLESS_OK."}],"source":{"kind":"user"},"role":"user","id":"8a0ac233-283c-4eb4-8bbd-5c50b7e99afe"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Run this advanced flow exactly once: define a host-only dynamic Cordis Package named Snapshot Marker; run and inspect snap-1/pkg-1 through run_code; delegate once to a direct spawn child; run one workflow that delegates to another spawn child; remove snap-1; then reply with exactly ADVANCED_HEADLESS_OK."}],"source":{"kind":"user"},"role":"user","id":"e203969f-330f-4886-8d34-82d0001db89b"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Run this advanced flow exactly","messageSeqs":[4],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model. Your working directory is {{cwd}}.\n\nVerify your work by running the code or tests. Keep answers brief and factual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\n# Dynamic Cordis Plugins\n\nDynamic Cordis plugins temporarily extend the current DSH process. A Plugin uses apply(ctx) to consume Services, listen to Events, provide Services, register model Tools, or register browser UI in Slots.\n\n- Plugin and Package definitions exist only in the current process. define itself does not modify repository source, configuration, or disk, and definitions do not survive a process restart.\n- The restricted execution environment prevents accidental misuse; it is not a security boundary for malicious code. Services obtained by dynamic code connect to the real runtime.\n\n## Make the user-facing plan clear first\n\n- Dynamic Cordis Plugins are one available implementation mechanism, not the default for every request. Consider whether one could help only when the user intends to design or create something, or when a temporary interface could materially aid the current work. The presence of these instructions or Tools, and discussion of Cordis itself, do not make a request a dynamic-Plugin task.\n- When Cordis is a plausible fit, infer the intended work target and lifetime from the request and conversation. Use it only when the outcome belongs to the current running harness and should be delivered as a temporary runtime extension. If that distinction is materially ambiguous, ask at most one concise question about the intended result or lifetime. Otherwise proceed with the matching workflow; do not require the user to know or choose Cordis as an implementation mechanism.\n- Once a dynamic Plugin is appropriate, decide whether the task creates a new Plugin or modifies the Plugin named by the user with @pluginId. Proceed directly when the goal is clear; do not ask for repeated confirmation.\n- Choose Host, Client, or both from the requested outcome. Do not propose a Client/browser UI when the task does not need visible page behavior, and do not avoid Client when the requested outcome is visual, interactive, or depends on page state. Host versus Client is an implementation choice; do not make the user choose it.\n- When a design direction or a potentially useful interface would materially affect the result, ask at most one concise outcome or creative-preference question and offer a few candidate directions. Otherwise proceed directly; do not conduct a multi-round interview or a complex questionnaire.\n- cordis_define only defines and presents code; it does not run it. After definition, explain the pluginId and packageId returned by the Host and whether the next step is a run or update.\n- cordis_run may require user approval. When it returns awaiting-approval, explain that the user must allow or reject it in the UI. Do not wait, retry, or claim that it is running.\n- When it returns starting, explain that the request has entered the asynchronous flow and the Client is still activating. starting does not mean success. Wait for the system to report the final result through steering context.\n- Do not request approval again after the user rejects it. After a technical failure, fix the same Plugin from its diagnostics; do not silently create a replacement Plugin.\n\n## Recommended workflow and Tools\n\nBefore creating, modifying, or repairing a Plugin, load the cordis-plugin-development Skill. The Skill provides requirement navigation, capability composition, complete examples, and troubleshooting. Treat Inspect Provider results as the source of truth for exact APIs.\n\n1. cordis_inspect_list: discover the current Host and Client Providers and their read-only query methods.\n2. cordis_inspect_query: use the returned platform, provider, method, and schema to query exact Service, Event, Builtin, Slot, Theme token, or Tool information.\n3. cordis_inspect_self: inspect the current Session's Plugins, Packages, version pointers, source, and diagnostics. Source is returned only when both pluginId and packageId are specified.\n4. cordis_define: create the first Package for a new Plugin or append an immutable Package to an existing Plugin. It defines code but does not run it.\n5. cordis_run: activate an exact Package. Use run for the first activation, restarting current, or rollback; use update to switch versions.\n6. cordis_stop: remove the current Run and pending approval request while retaining definitions, grants, and version pointers.\n7. cordis_undefine: permanently stop and delete a Plugin and all of its Packages. Use it only after confirming that the user no longer needs them.\n\n- Inspect and Catalog data only confirm capabilities, names, signatures, types, and registration protocols before code is written; they do not replace business APIs.\n- Query Service.listService and Event.listEvents without input to choose from their compact signature directories, then query the exact service or event before using it. Exact queries return the structured contract and only its referenced types.\n- At runtime, a Plugin must call real Services or listen to real Events. Do not cache, display, or depend on Inspect results as business data.\n\n## Identity, versions, and approval\n\n- pluginId identifies a Plugin that can be modified over time. For a new Plugin, submit only a semantic idPrefix of 3–6 lowercase English letters; the Host allocates the final ID.\n- packageId identifies one immutable Host/Client source version under a Plugin. To change code, define a new Package; never overwrite an old version.\n- pluginRunId identifies one activation attempt and connects its approval, Host/Client loading, private RPC, Run card, and errors.\n- currentPackageId is the most recent fully successful Package. Stopping, starting an update, or failing an update does not clear it.\n- nextPackageId is the target awaiting approval, being attempted, awaiting Client activation, or most recently failed.\n- A single check mark authorizes only the current Package; double check marks authorize future versions of the same Plugin. A grant remains in effect after a technical failure.\n- An update stops the old Run before starting the target Package. Failure does not automatically restart the old version; retry next with update or roll back to current with run.\n\nWhen the user enters @pluginId, the system injects identity, the default base Package, version pointers, and runtime status, but not source code:\n\n1. Call cordis_inspect_self(pluginId, packageId) to read the target source.\n2. Use cordis_define in existing mode to append a Package to the same Plugin.\n3. Call cordis_run in run or update mode according to the version relationship.\n\nNever silently create another Plugin for @pluginId. If the reference is unavailable because it was removed, belongs to another Session, or was lost on process restart, tell the user directly.\n\n## High-frequency errors that must be avoided\n\n### Services: ctx.get and inject\n\n- Read an optional Service with ctx.get('serviceName') by default and handle undefined.\n- Declare inject: ['serviceName'] on the returned Plugin object only when the Service is a hard dependency and the Plugin must enter waiting until Cordis reactivates it after the Service appears.\n- Read ctx.serviceName only after declaring that Service in inject. Never access an undeclared Service as a ctx property.\n\n```js\nreturn {\n inject: ['requiredService'],\n apply(ctx) {\n ctx.requiredService.someMethod()\n const optionalService = ctx.get('optionalService')\n if (optionalService !== undefined) optionalService.someMethod()\n },\n}\n```\n\n### Code: use plain JavaScript only\n\n- Host and Client code is not transformed by TypeScript, JSX, or a bundler.\n- Do not use TypeScript types, as, decorators, import, require, or JSX.\n- Client React code must use React.createElement(...); never write .\n- Do not assume that process, Buffer, window, document, fetch, native timers, or any other global is available. Query the corresponding platform's Builtins and Services first.\n\n### Data: do not serialize live data\n\n- Services, Events, Slots, Sessions, and their derived Cordis/DSH objects are internal live data, not ordinary JSON that can be dumped.\n- Do not apply JSON.stringify, structuredClone, recursive enumeration, full copying, or whole-object display to live data.\n- Read only the leaf fields required by the task, then construct the smallest owned data object without Host references.\n\n### Lifecycle: every side effect must be reversible\n\n- Services, Events, Tools, handlers, timers, Slots, styles, and theme overrides must all belong to the current Fiber.\n- Use ctx.effect(), ctx.on(), or official APIs that return a disposer so stop, update, or undefine removes every side effect.\n- The cordis-plugin-development Skill contains complete timer, Waterfall, Slot, theme, Tool, RPC, and React examples and troubleshooting guidance.\n\n## Host and Client\n\n- Host runs in the DSH Node.js process and is appropriate for files, networking, commands, Agent/Session access, Host Events, Services, model Tools, and JSON methods callable by the Client.\n- Client runs in the browser page and is appropriate for themes, layout, current page state, Tool cards, and Slot UI.\n- Host and Client communicate through Package-private JSON methods: Host uses harness.handle(method, handler), and Client uses host.call(method, args). The direction is Client→Host, and only lossless JSON may cross it.\n- Client UI must be registered in a queried Slot; apply() cannot directly return a React Element. Query Slots.listSubTree without root to choose from the compact purpose/topology tree, then query the exact root for its full registration contract and props before writing code.\n- See the Skill and Inspect Providers for Run-specific panels and exact Slot registration patterns.\n\n## Asynchronous results and recovery\n\n- Do not wait inside a Tool for approval or browser work that can happen only after the current turn ends.\n- Asynchronous success, rejection, and runtime errors update Run state and notify you through steering context.\n- After a technical failure, use cordis_inspect_self to read the exact Package source and its message/stack. Define a corrected Package under the same Plugin and retry autonomously.\n- Use the cordis-plugin-development Skill for other failure causes, repair procedures, and complete extension patterns.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.\n\n## Writing code for run_code\n\n`run_code` takes two required arguments: `code` — the body of an async TypeScript function (erasable syntax only — no `enum` or namespaces; type annotations are advisory, the code runs type-stripped) — and `description`, a short summary of what the program does. Inside the program:\n\n- Call tools as `await tools.name(args)` — quoted access for exotic names: `tools[\"my-tool\"](args)`. Every call resolves to the tool's typed canonical JSON value. Tool arguments must be lossless JSON.\n- A FAILED tool call rejects with `ToolCallError`, whose `toolName` identifies the failed tool and whose `message` is human-readable — `try/catch` it to handle and continue.\n- Independent read-only calls MAY overlap under `Promise.all` (safe calls run concurrently; mutating calls run alone, in submission order). Sequence dependent work with `await`.\n- Emit results with `return` and/or `console.log(...)`. ONLY what you print or return comes back to you — intermediate tool results never enter the conversation, so extract just what you need.\n\nThe available tools:\n\n```ts\ntype JsonValue = null | boolean | number | string | JsonValue[] | { [key: string]: JsonValue }\n\ninterface ToolArgsMap {\n /** Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`. */\n bash: {\n /** The bash command to execute. */\n command: string;\n /** Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\". */\n description: string;\n /** Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry. */\n timeoutMs?: number;\n /** Working directory for this command. Defaults to the session workspace; a relative path is resolved against it. */\n workdir?: string;\n /** Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies. */\n run_in_background?: boolean;\n } & Record;\n /** Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs. */\n cordis_define: {\n plugin: {\n kind: \"new\";\n /** Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix. */\n idPrefix: string;\n } | {\n kind: \"existing\";\n /** Exact ID of an existing Plugin; the new Package is appended to that instance. */\n pluginId: string;\n };\n /** Short, readable Package name. */\n name: string;\n /** One-sentence, user-facing description of the Package purpose. */\n purpose: string;\n code: {\n /** Plain JavaScript function body that returns the Host-half Cordis Plugin. */\n host?: string;\n /** Plain JavaScript function body that returns the browser Client-half Cordis Plugin. */\n client?: string;\n };\n } & Record;\n /** List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call. */\n cordis_inspect_list: Record;\n /** Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props. */\n cordis_inspect_query: {\n /** Runtime platform that owns the Provider. */\n platform: \"host\" | \"client\";\n /** Exact Provider ID returned by cordis_inspect_list. */\n provider: string;\n /** Exact method name declared by the Provider manifest. */\n method: string;\n /** Optional query input; it must satisfy the method input schema. */\n input?: JsonValue;\n } & Record;\n /** Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers. */\n cordis_inspect_self: {\n /** Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin. */\n pluginId?: string;\n /** Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned. */\n packageId?: string;\n } & Record;\n /** Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it. */\n cordis_run: {\n /** Stable Plugin ID returned by cordis_define. */\n pluginId: string;\n /** Exact immutable Package ID to activate under that Plugin. */\n packageId: string;\n /** Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package. */\n mode: \"run\" | \"update\";\n } & Record;\n /** Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal. */\n cordis_stop: {\n /** Stable dynamic Plugin ID to stop. */\n pluginId: string;\n } & Record;\n /** Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead. */\n cordis_undefine: {\n /** Stable dynamic Plugin ID to remove permanently. */\n pluginId: string;\n } & Record;\n /** Edit an existing UTF-8 text file by replacing literal text. */\n edit: {\n /** Path to edit, resolved by the filesystem backend. */\n file_path: string;\n /** Literal text to replace. Must match exactly. */\n old_string: string;\n /** Literal replacement text. Use an empty string to delete the match. */\n new_string: string;\n /** Replace all matches. Defaults to false; when false, old_string must appear exactly once. */\n replace_all?: boolean;\n } & Record;\n /** Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op. */\n interrupt_agent: {\n /** The agent id of the running agent to interrupt. */\n agent_id: string;\n } & Record;\n /** Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops. */\n job_kill: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Optional short reason, recorded in the log and forwarded to the job. */\n reason?: string;\n } & Record;\n /** List your background jobs (running and finished) with their ids, kinds, and statuses. */\n job_list: Record;\n /** Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap. */\n job_output: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive. */\n wait?: boolean;\n /** Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum. */\n timeout_ms?: number;\n } & Record;\n /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */\n ralph: {\n /** The immutable completion objective for every fresh Ralph round. */\n objective: string;\n /** Optional positive safe-integer round cap, bounded by the deployment ceiling. */\n maxRounds?: number;\n } & Record;\n /** Read a UTF-8 text file and return line-numbered content. */\n read: {\n /** Path to read, resolved by the filesystem backend. */\n file_path: string;\n /** 1-based first line to return. Defaults to 1. */\n offset?: number;\n /** Maximum number of lines to return. Defaults to 2000. */\n limit?: number;\n } & Record;\n /** Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered. */\n send_message: {\n /** The subagent id returned when the background subagent was started. */\n subagent_id: string;\n /** The message to deliver to the subagent. */\n message: string;\n } & Record;\n /** Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill. */\n skill: {\n /** The exact skill name from the available skills list. */\n name: string;\n } & Record;\n /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. */\n subagent: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */\n prompt: string;\n /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */\n run_in_background?: boolean;\n } & Record;\n /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result. */\n subagent_fork: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new. */\n prompt: string;\n } & Record;\n /** Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished). */\n todo_write: {\n /** The COMPLETE task list, replacing any previous list. */\n todos: ({\n /** What the task is — a short imperative line. */\n content: string;\n /** pending (not started) | in_progress (now) | completed (done). */\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n } & Record;\n /** Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn. The workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result. Script-body hooks: - `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly. - `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages. - `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`. - `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim. Misused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`. Constraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes. */\n workflow: {\n /** The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `). */\n script: string;\n /** The workflow identity block (plain JSON — never code). */\n meta: {\n /** Short kebab-case workflow name. */\n name: string;\n /** One-line description of what the workflow does. */\n description: string;\n /** Optional guidance on when this workflow applies. */\n whenToUse?: string;\n /** Optional phase declarations matched by phase() calls. */\n phases?: ({\n /** The phase title phase() calls match by exact string. */\n title: string;\n /** Optional one-line description of the phase. */\n detail?: string;\n /** Optional provider override this phase is expected to use. */\n provider?: string;\n /** Optional model override this phase is expected to use. */\n model?: string;\n } & Record)[];\n } & Record;\n /** Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}). */\n args?: Record;\n } & Record;\n /** Create or fully replace a UTF-8 text file. */\n write: {\n /** Path to write, resolved by the filesystem backend. */\n file_path: string;\n /** Full UTF-8 text content to write. */\n content: string;\n } & Record;\n}\n\ninterface ToolOutputMap {\n bash: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"foreground\";\n exitCode: number | null;\n signal: string | null;\n timedOut: boolean;\n aborted: boolean;\n timeoutMs: number;\n stdout: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n stderr: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n sandbox?: {\n mode: string;\n denied: boolean;\n enforcement?: string;\n runnerFailed?: boolean;\n };\n };\n cordis_define: {\n pluginId: string;\n packageId: string;\n name: string;\n purpose: string;\n hasHostHalf: boolean;\n hasClientHalf: boolean;\n };\n cordis_inspect_list: JsonValue;\n cordis_inspect_query: JsonValue;\n cordis_inspect_self: JsonValue;\n cordis_run: JsonValue;\n cordis_stop: {\n pluginId: string;\n };\n cordis_undefine: {\n pluginId: string;\n wasRunning: boolean;\n };\n edit: {\n path: string;\n before: string;\n after: string;\n };\n interrupt_agent: {\n accepted: boolean;\n };\n job_kill: {\n outcome: \"cancellation-requested\" | \"already-finished\";\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n job_list: ({\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n })[];\n job_output: {\n text: string;\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n ralph: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n read: {\n path: string;\n offset: number;\n lines: {\n number: number;\n text: string;\n }[];\n totalLines: number;\n };\n send_message: {\n messageId: string;\n };\n skill: {\n name: string;\n provider: string;\n resourceBase?: {\n kind: \"directory\";\n path: string;\n } | {\n kind: \"url\";\n url: string;\n } | {\n kind: \"opaque\";\n description: string;\n };\n content: string;\n };\n subagent: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n subagent_fork: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n todo_write: {\n todos: ({\n content: string;\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n counts: {\n pending: number;\n inProgress: number;\n completed: number;\n };\n };\n workflow: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n write: {\n path: string;\n operation: \"create\" | \"update\";\n before: string | null;\n after: string;\n };\n}\n\ntype ToolName = keyof ToolOutputMap\n\ndeclare class ToolCallError extends Error {\n readonly name: \"ToolCallError\";\n readonly toolName: ToolName;\n}\n\ndeclare const tools: {\n [K in ToolName]: (args: ToolArgsMap[K]) => Promise;\n}\n```","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"cordis_define","description":"Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs.","parameters":{"type":"object","properties":{"plugin":{"oneOf":[{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"new"},"idPrefix":{"type":"string","description":"Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix."}},"required":["kind","idPrefix"]},{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"existing"},"pluginId":{"type":"string","description":"Exact ID of an existing Plugin; the new Package is appended to that instance."}},"required":["kind","pluginId"]}]},"name":{"type":"string","description":"Short, readable Package name."},"purpose":{"type":"string","description":"One-sentence, user-facing description of the Package purpose."},"code":{"type":"object","additionalProperties":false,"properties":{"host":{"type":"string","description":"Plain JavaScript function body that returns the Host-half Cordis Plugin."},"client":{"type":"string","description":"Plain JavaScript function body that returns the browser Client-half Cordis Plugin."}}}},"required":["plugin","name","purpose","code"]}},{"name":"cordis_inspect_list","description":"List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call.","parameters":{"type":"object","properties":{}}},{"name":"cordis_inspect_query","description":"Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props.","parameters":{"type":"object","properties":{"platform":{"type":"string","description":"Runtime platform that owns the Provider.","enum":["host","client"]},"provider":{"type":"string","description":"Exact Provider ID returned by cordis_inspect_list."},"method":{"type":"string","description":"Exact method name declared by the Provider manifest."},"input":{"description":"Optional query input; it must satisfy the method input schema."}},"required":["platform","provider","method"]}},{"name":"cordis_inspect_self","description":"Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin."},"packageId":{"type":"string","description":"Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned."}}}},{"name":"cordis_run","description":"Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define."},"packageId":{"type":"string","description":"Exact immutable Package ID to activate under that Plugin."},"mode":{"type":"string","description":"Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package.","enum":["run","update"]}},"required":["pluginId","packageId","mode"]}},{"name":"cordis_stop","description":"Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to stop."}},"required":["pluginId"]}},{"name":"cordis_undefine","description":"Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to remove permanently."}},"required":["pluginId"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"run_code","description":"Execute a TypeScript program against the available tools. Takes two required arguments: `code`, the BODY of an async function (erasable syntax only; top-level `await` and `return` work), and `description`, a short summary of what the program does. Call tools as `await tools.name(args)` per the declarations in the system prompt. Only what you print or return comes back — curate it.","parameters":{"type":"object","properties":{"code":{"type":"string","description":"The program: the body of an async TypeScript function."},"description":{"type":"string","description":"Clear, concise description of what this program does in active voice, 5-10 words (shown in the UI). Examples: \"Count TODO markers across packages\"; \"Read failing test and its fixture\"; \"Rename config key in every cordis.yml\"."}},"required":["code","description"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model. Your working directory is {{cwd}}.\n\nVerify your work by running the code or tests. Keep answers brief and factual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\n# Dynamic Cordis Plugins\n\nDynamic Cordis plugins temporarily extend the current DSH process. A Plugin uses apply(ctx) to consume Services, listen to Events, provide Services, register model Tools, or register browser UI in Slots.\n\n- Plugin and Package definitions exist only in the current process. define itself does not modify repository source, configuration, or disk, and definitions do not survive a process restart.\n- The restricted execution environment prevents accidental misuse; it is not a security boundary for malicious code. Services obtained by dynamic code connect to the real runtime.\n\n## Make the user-facing plan clear first\n\n- Dynamic Cordis Plugins are one available implementation mechanism, not the default for every request. Consider whether one could help only when the user intends to design or create something, or when a temporary interface could materially aid the current work. The presence of these instructions or Tools, and discussion of Cordis itself, do not make a request a dynamic-Plugin task.\n- When Cordis is a plausible fit, infer the intended work target and lifetime from the request and conversation. Use it only when the outcome belongs to the current running harness and should be delivered as a temporary runtime extension. If that distinction is materially ambiguous, ask at most one concise question about the intended result or lifetime. Otherwise proceed with the matching workflow; do not require the user to know or choose Cordis as an implementation mechanism.\n- Once a dynamic Plugin is appropriate, decide whether the task creates a new Plugin or modifies the Plugin named by the user with @pluginId. Proceed directly when the goal is clear; do not ask for repeated confirmation.\n- Choose Host, Client, or both from the requested outcome. Do not propose a Client/browser UI when the task does not need visible page behavior, and do not avoid Client when the requested outcome is visual, interactive, or depends on page state. Host versus Client is an implementation choice; do not make the user choose it.\n- When a design direction or a potentially useful interface would materially affect the result, ask at most one concise outcome or creative-preference question and offer a few candidate directions. Otherwise proceed directly; do not conduct a multi-round interview or a complex questionnaire.\n- cordis_define only defines and presents code; it does not run it. After definition, explain the pluginId and packageId returned by the Host and whether the next step is a run or update.\n- cordis_run may require user approval. When it returns awaiting-approval, explain that the user must allow or reject it in the UI. Do not wait, retry, or claim that it is running.\n- When it returns starting, explain that the request has entered the asynchronous flow and the Client is still activating. starting does not mean success. Wait for the system to report the final result through steering context.\n- Do not request approval again after the user rejects it. After a technical failure, fix the same Plugin from its diagnostics; do not silently create a replacement Plugin.\n\n## Recommended workflow and Tools\n\nBefore creating, modifying, or repairing a Plugin, load the cordis-plugin-development Skill. The Skill provides requirement navigation, capability composition, complete examples, and troubleshooting. Treat Inspect Provider results as the source of truth for exact APIs.\n\n1. cordis_inspect_list: discover the current Host and Client Providers and their read-only query methods.\n2. cordis_inspect_query: use the returned platform, provider, method, and schema to query exact Service, Event, Builtin, Slot, Theme token, or Tool information.\n3. cordis_inspect_self: inspect the current Session's Plugins, Packages, version pointers, source, and diagnostics. Source is returned only when both pluginId and packageId are specified.\n4. cordis_define: create the first Package for a new Plugin or append an immutable Package to an existing Plugin. It defines code but does not run it.\n5. cordis_run: activate an exact Package. Use run for the first activation, restarting current, or rollback; use update to switch versions.\n6. cordis_stop: remove the current Run and pending approval request while retaining definitions, grants, and version pointers.\n7. cordis_undefine: permanently stop and delete a Plugin and all of its Packages. Use it only after confirming that the user no longer needs them.\n\n- Inspect and Catalog data only confirm capabilities, names, signatures, types, and registration protocols before code is written; they do not replace business APIs.\n- Query Service.listService and Event.listEvents without input to choose from their compact signature directories, then query the exact service or event before using it. Exact queries return the structured contract and only its referenced types.\n- At runtime, a Plugin must call real Services or listen to real Events. Do not cache, display, or depend on Inspect results as business data.\n\n## Identity, versions, and approval\n\n- pluginId identifies a Plugin that can be modified over time. For a new Plugin, submit only a semantic idPrefix of 3–6 lowercase English letters; the Host allocates the final ID.\n- packageId identifies one immutable Host/Client source version under a Plugin. To change code, define a new Package; never overwrite an old version.\n- pluginRunId identifies one activation attempt and connects its approval, Host/Client loading, private RPC, Run card, and errors.\n- currentPackageId is the most recent fully successful Package. Stopping, starting an update, or failing an update does not clear it.\n- nextPackageId is the target awaiting approval, being attempted, awaiting Client activation, or most recently failed.\n- A single check mark authorizes only the current Package; double check marks authorize future versions of the same Plugin. A grant remains in effect after a technical failure.\n- An update stops the old Run before starting the target Package. Failure does not automatically restart the old version; retry next with update or roll back to current with run.\n\nWhen the user enters @pluginId, the system injects identity, the default base Package, version pointers, and runtime status, but not source code:\n\n1. Call cordis_inspect_self(pluginId, packageId) to read the target source.\n2. Use cordis_define in existing mode to append a Package to the same Plugin.\n3. Call cordis_run in run or update mode according to the version relationship.\n\nNever silently create another Plugin for @pluginId. If the reference is unavailable because it was removed, belongs to another Session, or was lost on process restart, tell the user directly.\n\n## High-frequency errors that must be avoided\n\n### Services: ctx.get and inject\n\n- Read an optional Service with ctx.get('serviceName') by default and handle undefined.\n- Declare inject: ['serviceName'] on the returned Plugin object only when the Service is a hard dependency and the Plugin must enter waiting until Cordis reactivates it after the Service appears.\n- Read ctx.serviceName only after declaring that Service in inject. Never access an undeclared Service as a ctx property.\n\n```js\nreturn {\n inject: ['requiredService'],\n apply(ctx) {\n ctx.requiredService.someMethod()\n const optionalService = ctx.get('optionalService')\n if (optionalService !== undefined) optionalService.someMethod()\n },\n}\n```\n\n### Code: use plain JavaScript only\n\n- Host and Client code is not transformed by TypeScript, JSX, or a bundler.\n- Do not use TypeScript types, as, decorators, import, require, or JSX.\n- Client React code must use React.createElement(...); never write .\n- Do not assume that process, Buffer, window, document, fetch, native timers, or any other global is available. Query the corresponding platform's Builtins and Services first.\n\n### Data: do not serialize live data\n\n- Services, Events, Slots, Sessions, and their derived Cordis/DSH objects are internal live data, not ordinary JSON that can be dumped.\n- Do not apply JSON.stringify, structuredClone, recursive enumeration, full copying, or whole-object display to live data.\n- Read only the leaf fields required by the task, then construct the smallest owned data object without Host references.\n\n### Lifecycle: every side effect must be reversible\n\n- Services, Events, Tools, handlers, timers, Slots, styles, and theme overrides must all belong to the current Fiber.\n- Use ctx.effect(), ctx.on(), or official APIs that return a disposer so stop, update, or undefine removes every side effect.\n- The cordis-plugin-development Skill contains complete timer, Waterfall, Slot, theme, Tool, RPC, and React examples and troubleshooting guidance.\n\n## Host and Client\n\n- Host runs in the DSH Node.js process and is appropriate for files, networking, commands, Agent/Session access, Host Events, Services, model Tools, and JSON methods callable by the Client.\n- Client runs in the browser page and is appropriate for themes, layout, current page state, Tool cards, and Slot UI.\n- Host and Client communicate through Package-private JSON methods: Host uses harness.handle(method, handler), and Client uses host.call(method, args). The direction is Client→Host, and only lossless JSON may cross it.\n- Client UI must be registered in a queried Slot; apply() cannot directly return a React Element. Query Slots.listSubTree without root to choose from the compact purpose/topology tree, then query the exact root for its full registration contract and props before writing code.\n- See the Skill and Inspect Providers for Run-specific panels and exact Slot registration patterns.\n\n## Asynchronous results and recovery\n\n- Do not wait inside a Tool for approval or browser work that can happen only after the current turn ends.\n- Asynchronous success, rejection, and runtime errors update Run state and notify you through steering context.\n- After a technical failure, use cordis_inspect_self to read the exact Package source and its message/stack. Define a corrected Package under the same Plugin and retry autonomously.\n- Use the cordis-plugin-development Skill for other failure causes, repair procedures, and complete extension patterns.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.\n\n## Writing code for run_code\n\n`run_code` takes two required arguments: `code` — the body of an async TypeScript function (erasable syntax only — no `enum` or namespaces; type annotations are advisory, the code runs type-stripped) — and `description`, a short summary of what the program does. Inside the program:\n\n- Call tools as `await tools.name(args)` — quoted access for exotic names: `tools[\"my-tool\"](args)`. Every call resolves to the tool's typed canonical JSON value. Tool arguments must be lossless JSON.\n- A FAILED tool call rejects with `ToolCallError`, whose `toolName` identifies the failed tool and whose `message` is human-readable — `try/catch` it to handle and continue.\n- Independent read-only calls MAY overlap under `Promise.all` (safe calls run concurrently; mutating calls run alone, in submission order). Sequence dependent work with `await`.\n- Emit results with `return` and/or `console.log(...)`. Only what you print or return is program output. A successful tool result containing an image is attached after the run so you can inspect it on the next step; every other intermediate result stays out of the conversation, so extract just what you need.\n\nThe available tools:\n\n```ts\ntype JsonValue = null | boolean | number | string | JsonValue[] | { [key: string]: JsonValue }\n\ninterface ToolArgsMap {\n /** Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`. */\n bash: {\n /** The bash command to execute. */\n command: string;\n /** Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\". */\n description: string;\n /** Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry. */\n timeoutMs?: number;\n /** Working directory for this command. Defaults to the session workspace; a relative path is resolved against it. */\n workdir?: string;\n /** Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies. */\n run_in_background?: boolean;\n } & Record;\n /** Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs. */\n cordis_define: {\n plugin: {\n kind: \"new\";\n /** Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix. */\n idPrefix: string;\n } | {\n kind: \"existing\";\n /** Exact ID of an existing Plugin; the new Package is appended to that instance. */\n pluginId: string;\n };\n /** Short, readable Package name. */\n name: string;\n /** One-sentence, user-facing description of the Package purpose. */\n purpose: string;\n code: {\n /** Plain JavaScript function body that returns the Host-half Cordis Plugin. */\n host?: string;\n /** Plain JavaScript function body that returns the browser Client-half Cordis Plugin. */\n client?: string;\n };\n } & Record;\n /** List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call. */\n cordis_inspect_list: Record;\n /** Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props. */\n cordis_inspect_query: {\n /** Runtime platform that owns the Provider. */\n platform: \"host\" | \"client\";\n /** Exact Provider ID returned by cordis_inspect_list. */\n provider: string;\n /** Exact method name declared by the Provider manifest. */\n method: string;\n /** Optional query input; it must satisfy the method input schema. */\n input?: JsonValue;\n } & Record;\n /** Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers. */\n cordis_inspect_self: {\n /** Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin. */\n pluginId?: string;\n /** Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned. */\n packageId?: string;\n } & Record;\n /** Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it. */\n cordis_run: {\n /** Stable Plugin ID returned by cordis_define. */\n pluginId: string;\n /** Exact immutable Package ID to activate under that Plugin. */\n packageId: string;\n /** Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package. */\n mode: \"run\" | \"update\";\n } & Record;\n /** Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal. */\n cordis_stop: {\n /** Stable dynamic Plugin ID to stop. */\n pluginId: string;\n } & Record;\n /** Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead. */\n cordis_undefine: {\n /** Stable dynamic Plugin ID to remove permanently. */\n pluginId: string;\n } & Record;\n /** Edit an existing UTF-8 text file by replacing literal text. */\n edit: {\n /** Path to edit, resolved by the filesystem backend. */\n file_path: string;\n /** Literal text to replace. Must match exactly. */\n old_string: string;\n /** Literal replacement text. Use an empty string to delete the match. */\n new_string: string;\n /** Replace all matches. Defaults to false; when false, old_string must appear exactly once. */\n replace_all?: boolean;\n } & Record;\n /** Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op. */\n interrupt_agent: {\n /** The agent id of the running agent to interrupt. */\n agent_id: string;\n } & Record;\n /** Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops. */\n job_kill: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Optional short reason, recorded in the log and forwarded to the job. */\n reason?: string;\n } & Record;\n /** List your background jobs (running and finished) with their ids, kinds, and statuses. */\n job_list: Record;\n /** Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap. */\n job_output: {\n /** Job id returned by the tool that started the background work. */\n job_id: string;\n /** Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive. */\n wait?: boolean;\n /** Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum. */\n timeout_ms?: number;\n } & Record;\n /** Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields. */\n list_subagent_models: {\n /** Registered LLM provider id. Omit to list providers. */\n provider?: string;\n /** Exact model id to inspect. Requires provider; omit to list that provider's advertised models. */\n model?: string;\n } & Record;\n /** Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools. */\n ralph: {\n /** The immutable completion objective for every fresh Ralph round. */\n objective: string;\n /** Optional positive safe-integer round cap, bounded by the deployment ceiling. */\n maxRounds?: number;\n } & Record;\n /** Read a UTF-8 text file and return line-numbered content. */\n read: {\n /** Path to read, resolved by the filesystem backend. */\n file_path: string;\n /** 1-based first line to return. Defaults to 1. */\n offset?: number;\n /** Maximum number of lines to return. Defaults to 2000. */\n limit?: number;\n } & Record;\n /** Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered. */\n send_message: {\n /** The subagent id returned when the background subagent was started. */\n subagent_id: string;\n /** The message to deliver to the subagent. */\n message: string;\n } & Record;\n /** Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill. */\n skill: {\n /** The exact skill name from the available skills list. */\n name: string;\n } & Record;\n /** Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort. */\n subagent: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs. */\n prompt: string;\n /** LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route. */\n provider?: string;\n /** Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route. */\n model?: string;\n /** Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default. */\n reasoning_effort?: string;\n /** Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it. */\n run_in_background?: boolean;\n } & Record;\n /** Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result. */\n subagent_fork: {\n /** A short (3-5 word) description of the delegated task, for display. */\n description: string;\n /** The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new. */\n prompt: string;\n } & Record;\n /** Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished). */\n todo_write: {\n /** The COMPLETE task list, replacing any previous list. */\n todos: ({\n /** What the task is — a short imperative line. */\n content: string;\n /** pending (not started) | in_progress (now) | completed (done). */\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n } & Record;\n /** Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn. The workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result. Script-body hooks: - `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly. - `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages. - `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`. - `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim. Misused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`. Constraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes. */\n workflow: {\n /** The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `). */\n script: string;\n /** The workflow identity block (plain JSON — never code). */\n meta: {\n /** Short kebab-case workflow name. */\n name: string;\n /** One-line description of what the workflow does. */\n description: string;\n /** Optional guidance on when this workflow applies. */\n whenToUse?: string;\n /** Optional phase declarations matched by phase() calls. */\n phases?: ({\n /** The phase title phase() calls match by exact string. */\n title: string;\n /** Optional one-line description of the phase. */\n detail?: string;\n /** Optional provider override this phase is expected to use. */\n provider?: string;\n /** Optional model override this phase is expected to use. */\n model?: string;\n } & Record)[];\n } & Record;\n /** Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}). */\n args?: Record;\n } & Record;\n /** Create or fully replace a UTF-8 text file. */\n write: {\n /** Path to write, resolved by the filesystem backend. */\n file_path: string;\n /** Full UTF-8 text content to write. */\n content: string;\n } & Record;\n}\n\ninterface ToolOutputMap {\n bash: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"foreground\";\n exitCode: number | null;\n signal: string | null;\n timedOut: boolean;\n aborted: boolean;\n timeoutMs: number;\n stdout: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n stderr: {\n text: string;\n truncated: boolean;\n spillPath?: string;\n };\n sandbox?: {\n mode: string;\n denied: boolean;\n enforcement?: string;\n runnerFailed?: boolean;\n };\n };\n cordis_define: {\n pluginId: string;\n packageId: string;\n name: string;\n purpose: string;\n hasHostHalf: boolean;\n hasClientHalf: boolean;\n };\n cordis_inspect_list: JsonValue;\n cordis_inspect_query: JsonValue;\n cordis_inspect_self: JsonValue;\n cordis_run: JsonValue;\n cordis_stop: {\n pluginId: string;\n };\n cordis_undefine: {\n pluginId: string;\n wasRunning: boolean;\n };\n edit: {\n path: string;\n before: string;\n after: string;\n };\n interrupt_agent: {\n accepted: boolean;\n };\n job_kill: {\n outcome: \"cancellation-requested\" | \"already-finished\";\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n job_list: ({\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n })[];\n job_output: {\n text: string;\n job: {\n id: string;\n kind: string;\n label: string;\n status: \"running\" | \"stopping\" | \"completed\" | \"killed\" | \"failed\";\n detail?: string;\n startedAt: number;\n finishedAt?: number;\n };\n };\n list_subagent_models: string;\n ralph: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n read: {\n path: string;\n offset: number;\n lines: {\n number: number;\n text: string;\n }[];\n totalLines: number;\n };\n send_message: {\n messageId: string;\n };\n skill: {\n name: string;\n provider: string;\n resourceBase?: {\n kind: \"directory\";\n path: string;\n } | {\n kind: \"url\";\n url: string;\n } | {\n kind: \"opaque\";\n description: string;\n };\n content: string;\n };\n subagent: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n subagent_fork: {\n kind: \"background\";\n jobId: string;\n } | {\n kind: \"continuable\";\n subagentId: string;\n } | {\n kind: \"foreground\";\n runId: string;\n output: JsonValue[];\n };\n todo_write: {\n todos: ({\n content: string;\n status: \"pending\" | \"in_progress\" | \"completed\";\n })[];\n counts: {\n pending: number;\n inProgress: number;\n completed: number;\n };\n };\n workflow: {\n runId: string;\n agentsStarted: number;\n result: JsonValue;\n };\n write: {\n path: string;\n operation: \"create\" | \"update\";\n before: string | null;\n after: string;\n };\n}\n\ntype ToolName = keyof ToolOutputMap\n\ndeclare class ToolCallError extends Error {\n readonly name: \"ToolCallError\";\n readonly toolName: ToolName;\n}\n\ndeclare const tools: {\n [K in ToolName]: (args: ToolArgsMap[K]) => Promise;\n}\n```","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"cordis_define","description":"Define an immutable Cordis Package. For a new Plugin, use kind:\"new\" and provide only a semantic prefix of 3–6 lowercase English letters; the Host returns the final pluginId and packageId. To modify an existing Plugin, use kind:\"existing\" with its exact pluginId to append a Package without overwriting older versions. Provide at least one of code.host and code.client. Each value is a plain JavaScript function body that returns a Cordis Plugin; no TypeScript, JSX, or import transformation occurs. Query Inspect before depending on a Service, Event, Builtin, Slot, or token. Define only validates parameters and syntax and records source: it does not request approval, execute apply, or change currentPackageId. On success, call cordis_run with the returned IDs.","parameters":{"type":"object","properties":{"plugin":{"oneOf":[{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"new"},"idPrefix":{"type":"string","description":"Suggested semantic prefix of 3–6 lowercase English letters; the Host adds a unique numeric suffix."}},"required":["kind","idPrefix"]},{"type":"object","additionalProperties":false,"properties":{"kind":{"type":"string","const":"existing"},"pluginId":{"type":"string","description":"Exact ID of an existing Plugin; the new Package is appended to that instance."}},"required":["kind","pluginId"]}]},"name":{"type":"string","description":"Short, readable Package name."},"purpose":{"type":"string","description":"One-sentence, user-facing description of the Package purpose."},"code":{"type":"object","additionalProperties":false,"properties":{"host":{"type":"string","description":"Plain JavaScript function body that returns the Host-half Cordis Plugin."},"client":{"type":"string","description":"Plain JavaScript function body that returns the browser Client-half Cordis Plugin."}}}},"required":["plugin","name","purpose","code"]}},{"name":"cordis_inspect_list","description":"List every Cordis Inspect Provider currently known to the Host, including local Host Providers and the latest manifests synchronized from the Client. Each entry includes its platform, purpose, read-only methods, and input/output schemas. Call this Tool before creating or modifying a Package, then select the provider and method for cordis_inspect_query from its result. Do not guess names or treat an Inspect method as a business Service that Plugin code can call.","parameters":{"type":"object","properties":{}}},{"name":"cordis_inspect_query","description":"Run a read-only query explicitly declared by an Inspect Provider. platform, provider, and method must come from cordis_inspect_list, and input must satisfy that method's schema. Use this Tool before cordis_define to read exact Service methods, Event modes, Builtin signatures, Tool schemas, theme tokens, or live Slot trees and props. Host queries run locally. A Client query waits for the first valid page response and remains pending until a page answers or the Tool is cancelled. This Tool cannot invoke business Service methods or modify the runtime. For Service.listService and Event.listEvents, query without input to navigate the compact signature directory, then query the exact service or event for its structured contract and referenced types. For Slots.listSubTree, query without root to navigate the compact tree, then query the exact root for its complete registration contract and props.","parameters":{"type":"object","properties":{"platform":{"type":"string","description":"Runtime platform that owns the Provider.","enum":["host","client"]},"provider":{"type":"string","description":"Exact Provider ID returned by cordis_inspect_list."},"method":{"type":"string","description":"Exact method name declared by the Provider manifest."},"input":{"description":"Optional query input; it must satisfy the method input schema."}},"required":["platform","provider","method"]}},{"name":"cordis_inspect_self","description":"Inspect dynamic Cordis objects owned by the current Session at increasing levels of detail. With no IDs, list only Plugin summaries. With pluginId alone, return version pointers, the latest Run, and every Package summary. Only pluginId plus packageId returns that immutable Package's Host/Client source and runtime diagnostics. packageId cannot be supplied alone. Query an exact Package before handling @pluginId, repairing an asynchronous failure, or defining an updated version. This Tool is read-only: it neither executes code nor changes version pointers.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define or injected by @pluginId; omit it to list every current Plugin."},"packageId":{"type":"string","description":"Exact immutable Package ID owned by pluginId; when specified, source and diagnostics are returned."}}}},{"name":"cordis_run","description":"Activate one exact Package of a dynamic Plugin. Use mode:\"run\" for the first activation, restarting currentPackageId, or rollback. When current exists, use mode:\"update\" to switch to a different Package, even if the Plugin is currently stopped. An unauthorized Client Package creates an approval request and returns awaiting-approval; an authorized Package returns starting and continues asynchronously in the browser. Neither result waits for the final outcome inside the Tool. currentPackageId changes only after complete success; on failure, the old current and target next remain. Asynchronous success, rejection, or technical failure is reported through state and steering. After a technical failure, read diagnostics with cordis_inspect_self, correct the same Plugin, and retry autonomously. Do not request approval again after the user rejects it.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable Plugin ID returned by cordis_define."},"packageId":{"type":"string","description":"Exact immutable Package ID to activate under that Plugin."},"mode":{"type":"string","description":"Use run for the first activation, restarting current, or rollback; use update to switch from current to a different Package.","enum":["run","update"]}},"required":["pluginId","packageId","mode"]}},{"name":"cordis_stop","description":"Stop the current Run of a dynamic Plugin and cancel unfinished approval or activation requests. Retain the Plugin, every immutable Package, grants, currentPackageId, and nextPackageId so it can later run or update directly. Stopping an already stopped Plugin succeeds idempotently. Use this Tool to disable effects temporarily; use cordis_undefine for permanent removal.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to stop."}},"required":["pluginId"]}},{"name":"cordis_undefine","description":"Permanently remove a dynamic Plugin owned by the current Session. If it is running or awaiting approval, first stop it and cancel the request, then delete every Package, grant, and version pointer. After this returns, its pluginId, packageIds, @ reference, and Package business views are invalid; historical cards retain only a \"Plugin removed\" record. Do not call this Tool when versions must remain available for restart or rollback; use cordis_stop instead.","parameters":{"type":"object","properties":{"pluginId":{"type":"string","description":"Stable dynamic Plugin ID to remove permanently."}},"required":["pluginId"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"list_subagent_models","description":"Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.","parameters":{"type":"object","properties":{"provider":{"type":"string","description":"Registered LLM provider id. Omit to list providers."},"model":{"type":"string","description":"Exact model id to inspect. Requires provider; omit to list that provider's advertised models."}}}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"run_code","description":"Execute a TypeScript program against the available tools. Takes two required arguments: `code`, the BODY of an async function (erasable syntax only; top-level `await` and `return` work), and `description`, a short summary of what the program does. Call tools as `await tools.name(args)` per the declarations in the system prompt. Only what you print or return is program output — curate it. Image-bearing subtool results are attached after the run.","parameters":{"type":"object","properties":{"code":{"type":"string","description":"The program: the body of an async TypeScript function."},"description":{"type":"string","description":"Clear, concise description of what this program does in active voice, 5-10 words (shown in the UI). Examples: \"Count TODO markers across packages\"; \"Read failing test and its fixture\"; \"Rename config key in every cordis.yml\"."}},"required":["code","description"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"provider":{"type":"string","description":"LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."},"model":{"type":"string","description":"Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."},"reasoning_effort":{"type":"string","description":"Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"tool-call-delta","index":0,"id":"advanced-define","name":"cordis_define","argumentsDelta":"{\"plugin\":{\"kind\":\"new\",\"idPrefix\":\"snap\"},\"name\":\"Snapshot Marker\",\"purpose\":\"Exercise the dynamic Cordis Package lifecycle in the snapshot.\",\"code\":{\"host\":\"return { apply() {} }\"}}"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"advanced-define","name":"cordis_define","arguments":"{\"plugin\":{\"kind\":\"new\",\"idPrefix\":\"snap\"},\"name\":\"Snapshot Marker\",\"purpose\":\"Exercise the dynamic Cordis Package lifecycle in the snapshot.\",\"code\":{\"host\":\"return { apply() {} }\"}}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"usage","usage":{"inputTokens":3,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-define","name":"cordis_define","arguments":"{\"plugin\":{\"kind\":\"new\",\"idPrefix\":\"snap\"},\"name\":\"Snapshot Marker\",\"purpose\":\"Exercise the dynamic Cordis Package lifecycle in the snapshot.\",\"code\":{\"host\":\"return { apply() {} }\"}}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"e0351488-7bca-48d6-b7af-87d533858f47"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[8,9,10,11,12],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-define","name":"cordis_define","arguments":"{\"plugin\":{\"kind\":\"new\",\"idPrefix\":\"snap\"},\"name\":\"Snapshot Marker\",\"purpose\":\"Exercise the dynamic Cordis Package lifecycle in the snapshot.\",\"code\":{\"host\":\"return { apply() {} }\"}}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"749df5cd-1017-43c0-8c74-b2cab4752e59"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[8,9,10,11,12],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":1,"callId":"advanced-define","name":"cordis_define","arguments":"{\"plugin\":{\"kind\":\"new\",\"idPrefix\":\"snap\"},\"name\":\"Snapshot Marker\",\"purpose\":\"Exercise the dynamic Cordis Package lifecycle in the snapshot.\",\"code\":{\"host\":\"return { apply() {} }\"}}"}} -{"type":"tool/result","data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"advanced-define"},"content":[{"type":"tool-result","toolCallId":"advanced-define","content":[{"type":"text","text":"Defined snap-1/pkg-1 (Snapshot Marker); it is not running yet. Use cordis_run to activate this Package."}],"isError":false}],"role":"user","id":"9fd9ba62-1b6d-4bb7-98c6-97815e983026"},"meta":{"pluginId":"snap-1","packageId":"pkg-1"}},"sourceEventSeqs":[14],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"advanced-define"},"content":[{"type":"tool-result","toolCallId":"advanced-define","content":[{"type":"text","text":"Defined snap-1/pkg-1 (Snapshot Marker); it is not running yet. Use cordis_run to activate this Package."}],"isError":false}],"role":"user","id":"57e5798f-59ce-43ee-90c2-dcf4a66de1f3"},"meta":{"pluginId":"snap-1","packageId":"pkg-1"}},"sourceEventSeqs":[14],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":1}} {"type":"step/start","data":{"turn":1,"step":2}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -22,13 +22,13 @@ {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"advanced-code","name":"run_code","arguments":"{\"code\":\"const run = await tools.cordis_run({ pluginId: 'snap-1', packageId: 'pkg-1', mode: 'run' });\\nconst inspected = await tools.cordis_inspect_self({ pluginId: 'snap-1' });\\nreturn { run, inspected };\",\"description\":\"Run and inspect the dynamic Cordis Package\"}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"usage","usage":{"inputTokens":3,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":2,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-code","name":"run_code","arguments":"{\"code\":\"const run = await tools.cordis_run({ pluginId: 'snap-1', packageId: 'pkg-1', mode: 'run' });\\nconst inspected = await tools.cordis_inspect_self({ pluginId: 'snap-1' });\\nreturn { run, inspected };\",\"description\":\"Run and inspect the dynamic Cordis Package\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"6f6c3cc3-350b-4afb-a3d1-8f91b9494628"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[18,19,20,21,22],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":2,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-code","name":"run_code","arguments":"{\"code\":\"const run = await tools.cordis_run({ pluginId: 'snap-1', packageId: 'pkg-1', mode: 'run' });\\nconst inspected = await tools.cordis_inspect_self({ pluginId: 'snap-1' });\\nreturn { run, inspected };\",\"description\":\"Run and inspect the dynamic Cordis Package\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"a3df67b0-ba4f-4fc3-a359-05a63f636ba9"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[18,19,20,21,22],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":2,"callId":"advanced-code","name":"run_code","arguments":"{\"code\":\"const run = await tools.cordis_run({ pluginId: 'snap-1', packageId: 'pkg-1', mode: 'run' });\\nconst inspected = await tools.cordis_inspect_self({ pluginId: 'snap-1' });\\nreturn { run, inspected };\",\"description\":\"Run and inspect the dynamic Cordis Package\"}"}} {"type":"tool/code-dispatch-start","data":{"rootCallId":"advanced-code","parentCallId":"advanced-code","subCallId":"advanced-code:code:1","name":"cordis_run","arguments":{"pluginId":"snap-1","packageId":"pkg-1","mode":"run"}}} {"type":"tool/code-dispatch","data":{"rootCallId":"advanced-code","parentCallId":"advanced-code","subCallId":"advanced-code:code:1","name":"cordis_run","arguments":{"pluginId":"snap-1","packageId":"pkg-1","mode":"run"},"isError":false,"content":[{"type":"text","text":"snap-1/pkg-1 is running (run-1)."}]}} {"type":"tool/code-dispatch-start","data":{"rootCallId":"advanced-code","parentCallId":"advanced-code","subCallId":"advanced-code:code:2","name":"cordis_inspect_self","arguments":{"pluginId":"snap-1"}}} {"type":"tool/code-dispatch","data":{"rootCallId":"advanced-code","parentCallId":"advanced-code","subCallId":"advanced-code:code:2","name":"cordis_inspect_self","arguments":{"pluginId":"snap-1"},"isError":false,"content":[{"type":"text","text":"{\n \"mode\": \"plugin\",\n \"pluginId\": \"snap-1\",\n \"name\": \"Snapshot Marker\",\n \"packageCount\": 1,\n \"state\": \"running\",\n \"currentPackageId\": \"pkg-1\",\n \"activeRun\": {\n \"pluginRunId\": \"run-1\",\n \"packageId\": \"pkg-1\"\n },\n \"packages\": [\n {\n \"packageId\": \"pkg-1\",\n \"name\": \"Snapshot Marker\",\n \"purpose\": \"Exercise the dynamic Cordis Package lifecycle in the snapshot.\",\n \"hasHostHalf\": true,\n \"hasClientHalf\": false,\n \"isCurrent\": true,\n \"isNext\": false\n }\n ]\n}"}]}} -{"type":"tool/result","data":{"turn":1,"step":2,"message":{"source":{"kind":"tool","callId":"advanced-code"},"content":[{"type":"tool-result","toolCallId":"advanced-code","content":[{"type":"text","text":"{\n \"run\": {\n \"status\": \"running\",\n \"pluginId\": \"snap-1\",\n \"packageId\": \"pkg-1\",\n \"pluginRunId\": \"run-1\",\n \"currentPackageId\": \"pkg-1\",\n \"host\": {\n \"status\": \"running\",\n \"provides\": [],\n \"waitingFor\": []\n },\n \"client\": {\n \"status\": \"absent\",\n \"waitingFor\": []\n }\n },\n \"inspected\": {\n \"mode\": \"plugin\",\n \"pluginId\": \"snap-1\",\n \"name\": \"Snapshot Marker\",\n \"packageCount\": 1,\n \"state\": \"running\",\n \"currentPackageId\": \"pkg-1\",\n \"activeRun\": {\n \"pluginRunId\": \"run-1\",\n \"packageId\": \"pkg-1\"\n },\n \"packages\": [\n {\n \"packageId\": \"pkg-1\",\n \"name\": \"Snapshot Marker\",\n \"purpose\": \"Exercise the dynamic Cordis Package lifecycle in the snapshot.\",\n \"hasHostHalf\": true,\n \"hasClientHalf\": false,\n \"isCurrent\": true,\n \"isNext\": false\n }\n ]\n }\n}"}],"isError":false}],"role":"user","id":"40a17cbf-a853-4813-bbb5-7970cfbc7010"}},"sourceEventSeqs":[24],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":2,"message":{"source":{"kind":"tool","callId":"advanced-code"},"content":[{"type":"tool-result","toolCallId":"advanced-code","content":[{"type":"text","text":"{\n \"run\": {\n \"status\": \"running\",\n \"pluginId\": \"snap-1\",\n \"packageId\": \"pkg-1\",\n \"pluginRunId\": \"run-1\",\n \"currentPackageId\": \"pkg-1\",\n \"host\": {\n \"status\": \"running\",\n \"provides\": [],\n \"waitingFor\": []\n },\n \"client\": {\n \"status\": \"absent\",\n \"waitingFor\": []\n }\n },\n \"inspected\": {\n \"mode\": \"plugin\",\n \"pluginId\": \"snap-1\",\n \"name\": \"Snapshot Marker\",\n \"packageCount\": 1,\n \"state\": \"running\",\n \"currentPackageId\": \"pkg-1\",\n \"activeRun\": {\n \"pluginRunId\": \"run-1\",\n \"packageId\": \"pkg-1\"\n },\n \"packages\": [\n {\n \"packageId\": \"pkg-1\",\n \"name\": \"Snapshot Marker\",\n \"purpose\": \"Exercise the dynamic Cordis Package lifecycle in the snapshot.\",\n \"hasHostHalf\": true,\n \"hasClientHalf\": false,\n \"isCurrent\": true,\n \"isNext\": false\n }\n ]\n }\n}"}],"isError":false}],"role":"user","id":"2e591904-ffdd-4b86-824e-a4a5667b9383"}},"sourceEventSeqs":[24],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":2}} {"type":"step/start","data":{"turn":1,"step":3}} {"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -36,9 +36,9 @@ {"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"advanced-direct-child","name":"subagent","arguments":"{\"description\":\"Check direct child\",\"prompt\":\"Reply with exactly DIRECT_CHILD_OK and nothing else.\",\"run_in_background\":false}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"usage","usage":{"inputTokens":3,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":3,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-direct-child","name":"subagent","arguments":"{\"description\":\"Check direct child\",\"prompt\":\"Reply with exactly DIRECT_CHILD_OK and nothing else.\",\"run_in_background\":false}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"8bdea090-4a9a-451f-bb21-459af50472fa"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[32,33,34,35,36],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":3,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-direct-child","name":"subagent","arguments":"{\"description\":\"Check direct child\",\"prompt\":\"Reply with exactly DIRECT_CHILD_OK and nothing else.\",\"run_in_background\":false}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"93019032-9a3e-4947-aaff-4790cc42ecbb"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[32,33,34,35,36],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":3,"callId":"advanced-direct-child","name":"subagent","arguments":"{\"description\":\"Check direct child\",\"prompt\":\"Reply with exactly DIRECT_CHILD_OK and nothing else.\",\"run_in_background\":false}"}} -{"type":"tool/result","data":{"turn":1,"step":3,"message":{"source":{"kind":"tool","callId":"advanced-direct-child"},"content":[{"type":"tool-result","toolCallId":"advanced-direct-child","content":[{"type":"text","text":"DIRECT_CHILD_OK"}],"isError":false}],"role":"user","id":"c978c208-8ebd-4dd6-b997-606ca7de787e"}},"sourceEventSeqs":[38],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":3,"message":{"source":{"kind":"tool","callId":"advanced-direct-child"},"content":[{"type":"tool-result","toolCallId":"advanced-direct-child","content":[{"type":"text","text":"DIRECT_CHILD_OK"}],"isError":false}],"role":"user","id":"3b86b42c-4989-4490-a1f4-471dc8984c00"}},"sourceEventSeqs":[38],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":3}} {"type":"step/start","data":{"turn":1,"step":4}} {"type":"assistant/chunk","data":{"turn":1,"step":4,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -46,13 +46,13 @@ {"type":"assistant/chunk","data":{"turn":1,"step":4,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"advanced-workflow","name":"workflow","arguments":"{\"script\":\"phase('Delegate')\\nconst reply = await agent('Reply with exactly WORKFLOW_CHILD_OK and nothing else.', { label: 'workflow-child' })\\nreturn { reply }\",\"meta\":{\"name\":\"advanced-headless-snapshot\",\"description\":\"exercise one workflow child through the headless agent\"}}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":4,"chunk":{"type":"usage","usage":{"inputTokens":3,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":4,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":4,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-workflow","name":"workflow","arguments":"{\"script\":\"phase('Delegate')\\nconst reply = await agent('Reply with exactly WORKFLOW_CHILD_OK and nothing else.', { label: 'workflow-child' })\\nreturn { reply }\",\"meta\":{\"name\":\"advanced-headless-snapshot\",\"description\":\"exercise one workflow child through the headless agent\"}}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"e55b2b2e-497c-45d2-8115-16f317ae573f"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[42,43,44,45,46],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":4,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-workflow","name":"workflow","arguments":"{\"script\":\"phase('Delegate')\\nconst reply = await agent('Reply with exactly WORKFLOW_CHILD_OK and nothing else.', { label: 'workflow-child' })\\nreturn { reply }\",\"meta\":{\"name\":\"advanced-headless-snapshot\",\"description\":\"exercise one workflow child through the headless agent\"}}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"df7d2480-5bf2-4780-ab43-376d3f867b2a"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[42,43,44,45,46],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":4,"callId":"advanced-workflow","name":"workflow","arguments":"{\"script\":\"phase('Delegate')\\nconst reply = await agent('Reply with exactly WORKFLOW_CHILD_OK and nothing else.', { label: 'workflow-child' })\\nreturn { reply }\",\"meta\":{\"name\":\"advanced-headless-snapshot\",\"description\":\"exercise one workflow child through the headless agent\"}}"}} {"type":"tool-workflow/run-start","data":{"runId":"cd3d2666-94a1-4285-804e-c99630bc7b51","name":"advanced-headless-snapshot"}} {"type":"tool-workflow/agent-start","data":{"runId":"cd3d2666-94a1-4285-804e-c99630bc7b51","seq":1,"label":"workflow-child","phase":"Delegate","childId":"33333333-3333-4333-8333-333333333333"}} {"type":"tool-workflow/agent-end","data":{"runId":"cd3d2666-94a1-4285-804e-c99630bc7b51","seq":1,"outcome":"completed"}} {"type":"tool-workflow/run-end","data":{"runId":"cd3d2666-94a1-4285-804e-c99630bc7b51","stopReason":"completed"}} -{"type":"tool/result","data":{"turn":1,"step":4,"message":{"source":{"kind":"tool","callId":"advanced-workflow"},"content":[{"type":"tool-result","toolCallId":"advanced-workflow","content":[{"type":"text","text":"workflow \"advanced-headless-snapshot\" completed (1 agent).\nReturn value:\n{\n \"reply\": \"WORKFLOW_CHILD_OK\"\n}"}],"isError":false}],"role":"user","id":"f85f58fc-8d7e-4c9f-a0fe-caff480a9fec"}},"sourceEventSeqs":[48],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":4,"message":{"source":{"kind":"tool","callId":"advanced-workflow"},"content":[{"type":"tool-result","toolCallId":"advanced-workflow","content":[{"type":"text","text":"workflow \"advanced-headless-snapshot\" completed (1 agent).\nReturn value:\n{\n \"reply\": \"WORKFLOW_CHILD_OK\"\n}"}],"isError":false}],"role":"user","id":"9fd8c742-bba7-4a6a-9c91-ac5d19a77eff"}},"sourceEventSeqs":[48],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":4}} {"type":"step/start","data":{"turn":1,"step":5}} {"type":"assistant/chunk","data":{"turn":1,"step":5,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -60,9 +60,9 @@ {"type":"assistant/chunk","data":{"turn":1,"step":5,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"advanced-undefine","name":"cordis_undefine","arguments":"{\"pluginId\":\"snap-1\"}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":5,"chunk":{"type":"usage","usage":{"inputTokens":3,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":5,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":5,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-undefine","name":"cordis_undefine","arguments":"{\"pluginId\":\"snap-1\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"68bd993f-a0bd-4ea8-ad16-6b1a19e09bd3"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[56,57,58,59,60],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":5,"message":{"role":"assistant","content":[{"type":"tool-call","id":"advanced-undefine","name":"cordis_undefine","arguments":"{\"pluginId\":\"snap-1\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"ad845a3d-0053-4991-bccc-ba09bd640699"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[56,57,58,59,60],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":5,"callId":"advanced-undefine","name":"cordis_undefine","arguments":"{\"pluginId\":\"snap-1\"}"}} -{"type":"tool/result","data":{"turn":1,"step":5,"message":{"source":{"kind":"tool","callId":"advanced-undefine"},"content":[{"type":"tool-result","toolCallId":"advanced-undefine","content":[{"type":"text","text":"Removed dynamic Plugin snap-1 and all of its Packages."}],"isError":false}],"role":"user","id":"2d90e3b5-2a4c-4408-a1a0-3d009786a07b"}},"sourceEventSeqs":[62],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":5,"message":{"source":{"kind":"tool","callId":"advanced-undefine"},"content":[{"type":"tool-result","toolCallId":"advanced-undefine","content":[{"type":"text","text":"Removed dynamic Plugin snap-1 and all of its Packages."}],"isError":false}],"role":"user","id":"d21035a4-fcb1-46b5-af03-e9f9b7685444"}},"sourceEventSeqs":[62],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":5}} {"type":"step/start","data":{"turn":1,"step":6}} {"type":"assistant/chunk","data":{"turn":1,"step":6,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} @@ -70,6 +70,6 @@ {"type":"assistant/chunk","data":{"turn":1,"step":6,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"ADVANCED_HEADLESS_OK"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":6,"chunk":{"type":"usage","usage":{"inputTokens":3,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":6,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} -{"type":"assistant/message","data":{"turn":1,"step":6,"message":{"role":"assistant","content":[{"type":"text","text":"ADVANCED_HEADLESS_OK"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"f3823367-2e25-43d2-a129-70dc492b2a90"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[66,67,68,69,70],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":6,"message":{"role":"assistant","content":[{"type":"text","text":"ADVANCED_HEADLESS_OK"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"32f4ee19-e038-4c1c-9fc9-4288be6ca08f"},"usage":{"inputTokens":3,"outputTokens":3}},"sourceEventSeqs":[66,67,68,69,70],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":6}} {"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} diff --git a/examples/headless-agent/tests/snapshots/compaction-recovery/session.jsonl b/examples/headless-agent/tests/snapshots/compaction-recovery/session.jsonl index 575bbc5dfa..af15e860e3 100644 --- a/examples/headless-agent/tests/snapshots/compaction-recovery/session.jsonl +++ b/examples/headless-agent/tests/snapshots/compaction-recovery/session.jsonl @@ -1,32 +1,32 @@ {"type":"session","version":0,"id":"{{sessionId}}","createdAt":0,"cwd":"{{cwd}}","delegationDepth":0} -{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Establish a durable compaction premise before continuing. Record every part of this historical evidence: the snapshot uses keyless replay; persistence uses JSONL; the assembled headless application loads its real Cordis composition; model-visible inputs remain logged; tool calls and results remain paired and ordered; context overflow retains the original failure while recovery is attempted; compaction opens with compaction/start and closes with compaction/end; a successful auxiliary summary records compaction/summary provenance; the replacement surface shadows only an older balanced range; the checkpoint remains smaller than the history it replaces; the newest tool result remains verbatim; the retried request sees that checkpoint; the final response proves the same turn continued; deterministic snapshot evidence stays separate from the live-provider smoke; and no external API key is needed. Emit one alpha marker through bash, then finish the task after any required recovery with the exact words COMPACTION RECOVERED."}],"source":{"kind":"user"},"role":"user","id":"10eb2388-2d40-4564-af27-e7a5419fc14e"}]}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Establish a durable compaction premise before continuing. Record every part of this historical evidence: the snapshot uses keyless replay; persistence uses JSONL; the assembled headless application loads its real Cordis composition; model-visible inputs remain logged; tool calls and results remain paired and ordered; context overflow retains the original failure while recovery is attempted; compaction opens with compaction/start and closes with compaction/end; a successful auxiliary summary records compaction/summary provenance; the replacement surface shadows only an older balanced range; the checkpoint remains smaller than the history it replaces; the newest tool result remains verbatim; the retried request sees that checkpoint; the final response proves the same turn continued; deterministic snapshot evidence stays separate from the live-provider smoke; and no external API key is needed. Emit one alpha marker through bash, then finish the task after any required recovery with the exact words COMPACTION RECOVERED."}],"source":{"kind":"user"},"role":"user","id":"b187bcb5-4465-48d2-8aa1-ad12060ca048"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} {"type":"step/start","data":{"turn":1,"step":1}} -{"type":"user/message","data":{"content":[{"type":"text","text":"Establish a durable compaction premise before continuing. Record every part of this historical evidence: the snapshot uses keyless replay; persistence uses JSONL; the assembled headless application loads its real Cordis composition; model-visible inputs remain logged; tool calls and results remain paired and ordered; context overflow retains the original failure while recovery is attempted; compaction opens with compaction/start and closes with compaction/end; a successful auxiliary summary records compaction/summary provenance; the replacement surface shadows only an older balanced range; the checkpoint remains smaller than the history it replaces; the newest tool result remains verbatim; the retried request sees that checkpoint; the final response proves the same turn continued; deterministic snapshot evidence stays separate from the live-provider smoke; and no external API key is needed. Emit one alpha marker through bash, then finish the task after any required recovery with the exact words COMPACTION RECOVERED."}],"source":{"kind":"user"},"role":"user","id":"10eb2388-2d40-4564-af27-e7a5419fc14e"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Establish a durable compaction premise before continuing. Record every part of this historical evidence: the snapshot uses keyless replay; persistence uses JSONL; the assembled headless application loads its real Cordis composition; model-visible inputs remain logged; tool calls and results remain paired and ordered; context overflow retains the original failure while recovery is attempted; compaction opens with compaction/start and closes with compaction/end; a successful auxiliary summary records compaction/summary provenance; the replacement surface shadows only an older balanced range; the checkpoint remains smaller than the history it replaces; the newest tool result remains verbatim; the retried request sees that checkpoint; the final response proves the same turn continued; deterministic snapshot evidence stays separate from the live-provider smoke; and no external API key is needed. Emit one alpha marker through bash, then finish the task after any required recovery with the exact words COMPACTION RECOVERED."}],"source":{"kind":"user"},"role":"user","id":"b187bcb5-4465-48d2-8aa1-ad12060ca048"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Establish a durable compaction premise","messageSeqs":[4],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model.\n\nVerify your work by running the code or tests. Keep answers brief and\nfactual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model.\n\nVerify your work by running the code or tests. Keep answers brief and\nfactual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"list_subagent_models","description":"Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.","parameters":{"type":"object","properties":{"provider":{"type":"string","description":"Registered LLM provider id. Omit to list providers."},"model":{"type":"string","description":"Exact model id to inspect. Requires provider; omit to list that provider's advertised models."}}}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"provider":{"type":"string","description":"LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."},"model":{"type":"string","description":"Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."},"reasoning_effort":{"type":"string","description":"Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash","contextWindow":128000}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"tool-call-delta","index":0,"id":"call_compaction_marker","name":"bash","argumentsDelta":"{\"command\":\"printf 'alpha\\n'\",\"description\":\"Emit compaction premise marker\"}"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"call_compaction_marker","name":"bash","arguments":"{\"command\":\"printf 'alpha\\n'\",\"description\":\"Emit compaction premise marker\"}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"usage","usage":{"inputTokens":24,"outputTokens":6}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"call_compaction_marker","name":"bash","arguments":"{\"command\":\"printf 'alpha\\n'\",\"description\":\"Emit compaction premise marker\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"a71b2cfd-c18f-4a1b-82f6-e89fb371a87e"},"usage":{"inputTokens":24,"outputTokens":6}},"sourceEventSeqs":[8,9,10,11,12],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"call_compaction_marker","name":"bash","arguments":"{\"command\":\"printf 'alpha\\n'\",\"description\":\"Emit compaction premise marker\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"f6d534ce-c24d-49ee-ac3c-d05257fa03fb"},"usage":{"inputTokens":24,"outputTokens":6}},"sourceEventSeqs":[8,9,10,11,12],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":1,"callId":"call_compaction_marker","name":"bash","arguments":"{\"command\":\"printf 'alpha\\n'\",\"description\":\"Emit compaction premise marker\"}"}} -{"type":"tool/result","data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"call_compaction_marker"},"content":[{"type":"tool-result","toolCallId":"call_compaction_marker","content":[{"type":"text","text":"alpha\n"}],"isError":false}],"role":"user","id":"c9e68608-2dff-44bc-a344-b01006272378"}},"sourceEventSeqs":[14],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"call_compaction_marker"},"content":[{"type":"tool-result","toolCallId":"call_compaction_marker","content":[{"type":"text","text":"alpha\n"}],"isError":false}],"role":"user","id":"86b7cf93-1014-4334-a4af-5862c82fd5fb"}},"sourceEventSeqs":[14],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":1}} {"type":"step/start","data":{"turn":1,"step":2}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"finish","reason":{"kind":"error","failure":{"message":"snapshot request exceeded the model context window","code":"CONTEXT_WINDOW_EXCEEDED"}}}}} {"type":"compaction/start","data":{"compactionId":"338e88fa-e78b-4d49-bd38-8f919e85f5e1","turn":1}} {"type":"compaction/summary","data":{"compactionId":"338e88fa-e78b-4d49-bd38-8f919e85f5e1","summary":[{"type":"text","text":"The request established a durable compaction premise."}],"rawOutput":[{"type":"text","text":"The request established a durable compaction premise."}],"llmStreamCall":true,"shadowedRange":{"start":4,"end":4},"shadowedSeqs":[4],"shadowedTokenCount":266,"provider":"deepseek-official","model":"deepseek-v4-flash","maxTokens":32,"usage":{"inputTokens":20,"outputTokens":4}}} -{"type":"user/message","data":{"content":[{"type":"text","text":"This is an automatically generated checkpoint condensing an earlier span of the conversation to free up context. Treat the captured context as established background and build on it without restating it. Continue the task directly from the messages that follow, without acknowledging this checkpoint.\n\n"},{"type":"text","text":"The request established a durable compaction premise."},{"type":"text","text":""}],"source":{"kind":"plugin","plugin":"compact","compactionId":"338e88fa-e78b-4d49-bd38-8f919e85f5e1"},"role":"user","id":"3668b957-07a2-4cb7-96b1-98a23ac8cdb8"},"sourceEventSeqs":[19,20,4],"surfaceOp":{"op":"replace","start":4,"end":4}} +{"type":"user/message","data":{"content":[{"type":"text","text":"This is an automatically generated checkpoint condensing an earlier span of the conversation to free up context. Treat the captured context as established background and build on it without restating it. Continue the task directly from the messages that follow, without acknowledging this checkpoint.\n\n"},{"type":"text","text":"The request established a durable compaction premise."},{"type":"text","text":""}],"source":{"kind":"plugin","plugin":"compact","compactionId":"338e88fa-e78b-4d49-bd38-8f919e85f5e1"},"role":"user","id":"a4847ff0-6fa0-419b-8427-eade600c05ee"},"sourceEventSeqs":[19,20,4],"surfaceOp":{"op":"replace","start":4,"end":4}} {"type":"compaction/end","data":{"compactionId":"338e88fa-e78b-4d49-bd38-8f919e85f5e1","turn":1}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"text-delta","index":0,"text":"COMPACTION RECOVERED"}}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"COMPACTION RECOVERED"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"usage","usage":{"inputTokens":20,"outputTokens":4}}}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} -{"type":"assistant/message","data":{"turn":1,"step":2,"message":{"role":"assistant","content":[{"type":"text","text":"COMPACTION RECOVERED"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"bcbfd4ff-60e5-4634-ae39-4de3708a8abc"},"usage":{"inputTokens":20,"outputTokens":4}},"sourceEventSeqs":[23,24,25,26,27],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":2,"message":{"role":"assistant","content":[{"type":"text","text":"COMPACTION RECOVERED"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"18a03939-b8df-491e-ac8a-a74938f2aa4c"},"usage":{"inputTokens":20,"outputTokens":4}},"sourceEventSeqs":[23,24,25,26,27],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":2}} {"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} diff --git a/examples/headless-agent/tests/snapshots/pty-tools/session.jsonl b/examples/headless-agent/tests/snapshots/pty-tools/session.jsonl index 6e5148b4de..a3b7e4dc17 100644 --- a/examples/headless-agent/tests/snapshots/pty-tools/session.jsonl +++ b/examples/headless-agent/tests/snapshots/pty-tools/session.jsonl @@ -1,21 +1,21 @@ {"type":"session","version":0,"id":"{{sessionId}}","createdAt":0,"cwd":"{{cwd}}","delegationDepth":0} -{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Exercise the six PTY tools in order, including one missing-session signal error, then reply DONE."}],"source":{"kind":"user"},"role":"user","id":"d35cdacd-b5e6-4968-b7a3-5ec48f403ef7"}]}} +{"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Exercise the six PTY tools in order, including one missing-session signal error, then reply DONE."}],"source":{"kind":"user"},"role":"user","id":"41d289fe-fc8a-464a-b3e1-db2c0189cae9"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} {"type":"step/start","data":{"turn":1,"step":1}} -{"type":"user/message","data":{"content":[{"type":"text","text":"Exercise the six PTY tools in order, including one missing-session signal error, then reply DONE."}],"source":{"kind":"user"},"role":"user","id":"d35cdacd-b5e6-4968-b7a3-5ec48f403ef7"},"surfaceOp":"append"} -{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."}]},"role":"user","id":"053af702-9950-4860-913a-3c7e45a54f9d"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Exercise the six PTY tools in order, including one missing-session signal error, then reply DONE."}],"source":{"kind":"user"},"role":"user","id":"41d289fe-fc8a-464a-b3e1-db2c0189cae9"},"surfaceOp":"append"} +{"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."}]},"role":"user","id":"21b5e1f2-ef74-44cf-820b-d2a0bf77a519"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Exercise the six PTY tools","messageSeqs":[4],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model.\n\nVerify your work by running the code or tests. Keep answers brief and\nfactual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nUse a terminal session only when work needs persistent terminal state or interactive stdin; prefer shell/read/write/edit for bounded one-shot operations. Track every terminal session id and close sessions that no longer matter. An inferred_idle or timeout result does not prove the foreground command exited.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"terminal_close","description":"Close one persistent terminal and wait until its captured owned process tree is gone.","parameters":{"type":"object","properties":{"sessionId":{"type":"string","description":"Terminal session id."}},"required":["sessionId"]}},{"name":"terminal_list","description":"List persistent terminal sessions owned by the current agent.","parameters":{"type":"object","properties":{}}},{"name":"terminal_open","description":"Create a persistent, owner-isolated terminal session from a registered backend type. Use this for shell or REPL state that must survive across tool calls.","parameters":{"type":"object","properties":{"type":{"type":"string","description":"Registered terminal backend type, usually \"shell\"."},"name":{"type":"string","description":"Optional owner-local display name such as \"main\" or \"gdb\"."},"cwd":{"type":"string","description":"Initial working directory. Defaults to the deployment workspace root."}},"required":["type"]}},{"name":"terminal_read","description":"Read a bounded page of retained output from a persistent terminal without sending input.","parameters":{"type":"object","properties":{"sessionId":{"type":"string","description":"Terminal session id."},"offset":{"type":"number","description":"Newest-relative line offset (default 0)."},"count":{"type":"number","description":"Requested line count (default 500; backend caps apply)."}},"required":["sessionId"]}},{"name":"terminal_send","description":"Send text to a persistent terminal. By default Enter is submitted and the call waits for a prompt, stdin wait, output silence, timeout, or session exit. Background mode returns a job id for job_output/job_kill.","parameters":{"type":"object","properties":{"sessionId":{"type":"string","description":"Terminal session id returned by terminal_open or terminal_list."},"text":{"type":"string","description":"UTF-8 text to write to the terminal."},"submit":{"type":"boolean","description":"Submit Enter after text (default true). Set false for control characters or incomplete REPL input."},"run_in_background":{"type":"boolean","description":"Return a job id immediately; collect with job_output or stop with job_kill."}},"required":["sessionId","text"]}},{"name":"terminal_signal","description":"Send an allowed signal to the current foreground process group of a persistent terminal.","parameters":{"type":"object","properties":{"sessionId":{"type":"string","description":"Terminal session id."},"signal":{"type":"string","description":"Signal to deliver. Shell-targeted SIGKILL is rejected; use terminal_close.","enum":["SIGINT","SIGTERM","SIGKILL","SIGTSTP","SIGHUP"]}},"required":["sessionId","signal"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"You are an AI agent powered by DeepSeek Harness.\n\nYou are headless-agent, a coding assistant powered by the deepseek-v4-flash model.\n\nVerify your work by running the code or tests. Keep answers brief and\nfactual.\n\n\nUse the read tool — not shell commands like cat — to inspect text files. Results include line numbers. Use offset and limit to continue reading large files.\n\nUse the write tool to create files or completely replace file contents. Existing files are overwritten, so read an existing file first (the default fs-observation-policy requires it) and prefer edit for targeted changes.\n\nUse the edit tool for targeted changes to existing UTF-8 text files. It replaces literal old_string with new_string; by default old_string must appear exactly once. If old_string appears multiple times, provide a more specific old_string or set replace_all to true. Read the file first (the default fs-observation-policy requires it), unless you just created or edited it in this session.\n\nCheck the [exit code: N] marker on every bash result; investigate failures before moving on.\n\nUse a terminal session only when work needs persistent terminal state or interactive stdin; prefer shell/read/write/edit for bounded one-shot operations. Track every terminal session id and close sessions that no longer matter. An inferred_idle or timeout result does not prove the foreground command exited.\n\nTrack every background job id you start. You are notified in-session when a job finishes — do not busy-poll or sleep on one; keep working on independent steps and do not duplicate a running job's work. Before giving a final answer, collect every still-relevant job with job_output (set wait: true only when you are genuinely blocked on it), and job_kill jobs that stopped mattering.\n\nUse the workflow tool ONLY when the user explicitly asks for a workflow or for large multi-agent orchestration: you write a JavaScript script (the tool description documents the exact format) that fans work out across many subagents with phases and structured results. For one or two delegations, prefer plain subagent calls.\n\nUse the ralph tool ONLY when the direct human explicitly asks for a Ralph loop or fresh-agent iterative execution. Each Ralph round starts a fresh child with no conversation seed and uses the shared workspace as durable memory. Completion and blockers are worker reports, not independent evaluation. Use same-session goal tools for ordinary long-running objectives, and plain subagents or workflows for bounded delegation and fan-out.\n\nUse subagent in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set `run_in_background: false` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.","tools":[{"name":"bash","description":"Execute a bash command (`bash -c`) and return its stdout/stderr. Each call runs in a fresh shell: no state (cwd, variables, functions) persists between calls — pass `workdir` instead of using `cd`. Non-zero exits are reported as `[exit code: N]`. Current harness environment facts are exposed through managed `$DSH_*` variables; inspect them when needed. Commands may run under a file sandbox; a blocked file operation is reported as `[sandbox: file access denied under mode]` — a policy denial, not a bug in the command; do not retry another way. Long output is truncated to its tail; the full output is saved to a file whose path is reported when available. Set `run_in_background: true` for long-running commands: the call returns a job id immediately; read its output with `job_output` and stop it with `job_kill`.","parameters":{"type":"object","properties":{"command":{"type":"string","description":"The bash command to execute."},"description":{"type":"string","description":"Clear, concise description of what this command does in active voice, 5-10 words (shown in the UI). Examples: \"ls\" → \"List files in current directory\"; \"git status\" → \"Show working tree status\"; \"npm install\" → \"Install package dependencies\"."},"timeoutMs":{"type":"number","description":"Timeout in milliseconds. The executor applies its configured default and cap, and kills the command on expiry."},"workdir":{"type":"string","description":"Working directory for this command. Defaults to the session workspace; a relative path is resolved against it."},"run_in_background":{"type":"boolean","description":"Run in the background and return a job id immediately (collect with job_output, stop with job_kill). No timeout applies."}},"required":["command","description"]}},{"name":"edit","description":"Edit an existing UTF-8 text file by replacing literal text.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to edit, resolved by the filesystem backend."},"old_string":{"type":"string","description":"Literal text to replace. Must match exactly."},"new_string":{"type":"string","description":"Literal replacement text. Use an empty string to delete the match."},"replace_all":{"type":"boolean","description":"Replace all matches. Defaults to false; when false, old_string must appear exactly once."}},"required":["file_path","old_string","new_string"]}},{"name":"interrupt_agent","description":"Request cancellation of a background agent's current turn by its agent id. The target may be your direct child or a deeper agent created under you. Only the current turn stops: messages already queued for the agent stay parked until a later send_message, agents it started keep running, and the agent itself stays available for follow-ups. This call returns as soon as the stop request is accepted, so the target may keep running briefly; interrupting an agent that already finished is an accepted no-op.","parameters":{"type":"object","properties":{"agent_id":{"type":"string","description":"The agent id of the running agent to interrupt."}},"required":["agent_id"]}},{"name":"job_kill","description":"Request cancellation of a running background job by job id. Returns immediately; the job settles as killed once its work actually stops.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"reason":{"type":"string","description":"Optional short reason, recorded in the log and forwarded to the job."}},"required":["job_id"]}},{"name":"job_list","description":"List your background jobs (running and finished) with their ids, kinds, and statuses.","parameters":{"type":"object","properties":{}}},{"name":"job_output","description":"Read a background job. Stream jobs return only output since the previous read; final-output jobs return their result after settlement. Every response ends with `[status: ...]`. Reads are non-blocking unless `wait: true`, which waits up to the configured cap.","parameters":{"type":"object","properties":{"job_id":{"type":"string","description":"Job id returned by the tool that started the background work."},"wait":{"type":"boolean","description":"Block until the job reaches a terminal status or the timeout expires. A timed-out wait returns [status: running] and leaves the job alive."},"timeout_ms":{"type":"number","description":"Max wait in milliseconds (only meaningful with wait: true). Defaults to the configured wait timeout; capped by the configured maximum."}},"required":["job_id"]}},{"name":"list_subagent_models","description":"Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list registered providers, with `provider` to list its advertised models, or with `provider` and `model` to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may accept an unlisted model id. Use the returned ids with a delegation tool's `provider`, `model`, and `reasoning_effort` fields.","parameters":{"type":"object","properties":{"provider":{"type":"string","description":"Registered LLM provider id. Omit to list providers."},"model":{"type":"string","description":"Exact model id to inspect. Requires provider; omit to list that provider's advertised models."}}}},{"name":"ralph","description":"Run a foreground fresh-agent Ralph loop toward one immutable objective. Use only when the direct human explicitly asks for Ralph or fresh-agent iteration. Each round opens a new child with no parent conversation or prior child session; the shared workspace is long-term memory, and only a bounded structured report crosses rounds. The call returns when a worker reports completion or a concrete blocker, or at the round limit. Ordinary long-running same-session work belongs to goal tools.","parameters":{"type":"object","properties":{"objective":{"type":"string","description":"The immutable completion objective for every fresh Ralph round."},"maxRounds":{"type":"number","description":"Optional positive safe-integer round cap, bounded by the deployment ceiling."}},"required":["objective"]}},{"name":"read","description":"Read a UTF-8 text file and return line-numbered content.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to read, resolved by the filesystem backend."},"offset":{"type":"number","description":"1-based first line to return. Defaults to 1."},"limit":{"type":"number","description":"Maximum number of lines to return. Defaults to 2000."}},"required":["file_path"]}},{"name":"send_message","description":"Send a message to a background subagent by its subagent id, continuing the same conversation. It becomes the subagent's next turn: if it is still working, the message waits until its current turn finishes, so it cannot redirect work already underway. This call returns no answer from the subagent — only confirmation that the message was delivered — so use it to give it more work. A failure means the message was NOT delivered.","parameters":{"type":"object","properties":{"subagent_id":{"type":"string","description":"The subagent id returned when the background subagent was started."},"message":{"type":"string","description":"The message to deliver to the subagent."}},"required":["subagent_id","message"]}},{"name":"skill","description":"Load the full instructions for an available skill. Call this with the exact skill name from the session skill catalog before acting on a task that names or clearly matches that skill.","parameters":{"type":"object","properties":{"name":{"type":"string","description":"The exact skill name from the available skills list."}},"required":["name"]}},{"name":"subagent","description":"Delegate a self-contained task to a subagent (a separate agent that works in its own context) to offload focused, independent work — research, a scoped implementation, an analysis — so it does not consume this conversation's context. The subagent returns its result, not its intermediate steps. Give it a complete, standalone prompt: it does not see this conversation. This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result. Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model's default effort.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The complete, self-contained task for the subagent. It does not share this conversation's context, so include everything it needs."},"provider":{"type":"string","description":"LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route."},"model":{"type":"string","description":"Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route."},"reasoning_effort":{"type":"string","description":"Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model's default."},"run_in_background":{"type":"boolean","description":"Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it."}},"required":["description","prompt"]}},{"name":"subagent_fork","description":"Delegate a task to a subagent that inherits this conversation: a child agent seeded with all completed turns so far (it does not see the current in-flight turn). Use this when the subtask builds on this conversation's context — a follow-up analysis, a review, a continuation — without consuming this conversation's context for the work itself. You receive its result, not its intermediate steps. This call waits for the subagent and returns its result.","parameters":{"type":"object","properties":{"description":{"type":"string","description":"A short (3-5 word) description of the delegated task, for display."},"prompt":{"type":"string","description":"The task for the subagent. It already sees this conversation's completed turns, so build on them freely and state only what is new."}},"required":["description","prompt"]}},{"name":"terminal_close","description":"Close one persistent terminal and wait until its captured owned process tree is gone.","parameters":{"type":"object","properties":{"sessionId":{"type":"string","description":"Terminal session id."}},"required":["sessionId"]}},{"name":"terminal_list","description":"List persistent terminal sessions owned by the current agent.","parameters":{"type":"object","properties":{}}},{"name":"terminal_open","description":"Create a persistent, owner-isolated terminal session from a registered backend type. Use this for shell or REPL state that must survive across tool calls.","parameters":{"type":"object","properties":{"type":{"type":"string","description":"Registered terminal backend type, usually \"shell\"."},"name":{"type":"string","description":"Optional owner-local display name such as \"main\" or \"gdb\"."},"cwd":{"type":"string","description":"Initial working directory. Defaults to the deployment workspace root."}},"required":["type"]}},{"name":"terminal_read","description":"Read a bounded page of retained output from a persistent terminal without sending input.","parameters":{"type":"object","properties":{"sessionId":{"type":"string","description":"Terminal session id."},"offset":{"type":"number","description":"Newest-relative line offset (default 0)."},"count":{"type":"number","description":"Requested line count (default 500; backend caps apply)."}},"required":["sessionId"]}},{"name":"terminal_send","description":"Send text to a persistent terminal. By default Enter is submitted and the call waits for a prompt, stdin wait, output silence, timeout, or session exit. Background mode returns a job id for job_output/job_kill.","parameters":{"type":"object","properties":{"sessionId":{"type":"string","description":"Terminal session id returned by terminal_open or terminal_list."},"text":{"type":"string","description":"UTF-8 text to write to the terminal."},"submit":{"type":"boolean","description":"Submit Enter after text (default true). Set false for control characters or incomplete REPL input."},"run_in_background":{"type":"boolean","description":"Return a job id immediately; collect with job_output or stop with job_kill."}},"required":["sessionId","text"]}},{"name":"terminal_signal","description":"Send an allowed signal to the current foreground process group of a persistent terminal.","parameters":{"type":"object","properties":{"sessionId":{"type":"string","description":"Terminal session id."},"signal":{"type":"string","description":"Signal to deliver. Shell-targeted SIGKILL is rejected; use terminal_close.","enum":["SIGINT","SIGTERM","SIGKILL","SIGTSTP","SIGHUP"]}},"required":["sessionId","signal"]}},{"name":"todo_write","description":"Record and update a structured task list for the current work. Send the ENTIRE list every call — it REPLACES the previous list (there are no partial updates, no per-item edits). Use it to plan multi-step work and show progress: add one todo per concrete step before you start. Mark every todo being actively worked on `in_progress` — several at once when work genuinely runs in parallel (e.g. concurrent subagents or background commands), one for sequential work; while work remains, at least one task should be `in_progress`. Mark a todo `completed` the moment it is done (do not batch completions), and allow no `in_progress` item only once all work is complete. Skip the list for trivial single-step tasks. Statuses: `pending` (not started), `in_progress` (being worked on now), `completed` (finished).","parameters":{"type":"object","properties":{"todos":{"type":"array","description":"The COMPLETE task list, replacing any previous list.","items":{"type":"object","additionalProperties":false,"properties":{"content":{"type":"string","description":"What the task is — a short imperative line."},"status":{"type":"string","description":"pending (not started) | in_progress (now) | completed (done).","enum":["pending","in_progress","completed"]}},"required":["content","status"]}}},"required":["todos"]}},{"name":"workflow","description":"Run a JavaScript workflow script that orchestrates subagents at scale. Use this for work that fans out across many independent pieces — an audit over many files, a migration, multi-angle research, adversarial verification of findings — where you write the orchestration as a script instead of delegating turn by turn.\n\nThe workflow's identity rides the `meta` parameter as JSON: required `name` (short kebab-case) and `description` strings, optional `whenToUse` string and `phases` array (`{title, detail?, provider?, model?}`). The `script` parameter is the plain JavaScript body ONLY (NOT TypeScript, and NO `export const meta` statement — meta is a parameter, not code), running with top-level await; end with `return ` — the value must be JSON-serializable and is this tool's result.\n\nScript-body hooks:\n- `agent(prompt, opts?): Promise` — run one subagent to completion. Without `opts.schema` it resolves to the child's final text; with `opts.schema` (an object-rooted JSON Schema using ONLY type/properties/required/additionalProperties/items/enum/const/oneOf — no pattern/format/numeric bounds) it resolves to the validated object. Resolves `null` when the child fails (filter with `.filter(Boolean)`). Other opts: `label` (display), `phase` (progress group), and independent `provider`/`model` LLM target overrides (either may be provided alone). Anything else (`effort`/`isolation`/`agentType`) is rejected loudly.\n- `pipeline(items, ...stages): Promise` — run each item through the stages independently with NO barrier between stages (prefer this for multi-stage work). Each stage receives `(prev, item, index)`. An ordinary stage throw drops that ITEM to `null` and skips its remaining stages.\n- `parallel(thunks): Promise` — run zero-argument functions concurrently and await ALL of them (a barrier; use only when a stage genuinely needs every prior result together). A throwing thunk resolves to `null`.\n- `phase(title)` — start a progress phase; `log(message)` — narrate progress; `args` — the tool call's `args` input, verbatim.\n\nMisused hooks (bad arguments, unknown options, unsupported schemas, tripped caps) throw errors that ALWAYS kill the script — they never dissolve into a per-item `null`.\n\nConstraints: concurrency and total-agent caps apply; no filesystem, network, timers, or Node.js APIs are provided — the agents do the work, the script only coordinates them. The run executes in the foreground: this call returns when the whole script finishes.","parameters":{"type":"object","properties":{"script":{"type":"string","description":"The plain-JS workflow script body (top-level await allowed; NO `export const meta` statement; end with `return `)."},"meta":{"type":"object","description":"The workflow identity block (plain JSON — never code).","additionalProperties":true,"properties":{"name":{"type":"string","description":"Short kebab-case workflow name."},"description":{"type":"string","description":"One-line description of what the workflow does."},"whenToUse":{"type":"string","description":"Optional guidance on when this workflow applies."},"phases":{"type":"array","description":"Optional phase declarations matched by phase() calls.","items":{"type":"object","additionalProperties":true,"properties":{"title":{"type":"string","description":"The phase title phase() calls match by exact string."},"detail":{"type":"string","description":"Optional one-line description of the phase."},"provider":{"type":"string","description":"Optional provider override this phase is expected to use."},"model":{"type":"string","description":"Optional model override this phase is expected to use."}},"required":["title"]}}},"required":["name","description"]},"args":{"type":"object","description":"Optional JSON input exposed to the script as the `args` global (wrap a bare list as a field, e.g. {\"files\": [...]}).","additionalProperties":true}},"required":["script","meta"]}},{"name":"write","description":"Create or fully replace a UTF-8 text file.","parameters":{"type":"object","properties":{"file_path":{"type":"string","description":"Path to write, resolved by the filesystem backend."},"content":{"type":"string","description":"Full UTF-8 text content to write."}},"required":["file_path","content"]}}]},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"tool-call-delta","index":0,"id":"pty-spawn","name":"terminal_open","argumentsDelta":"{\"type\":\"shell\",\"name\":\"main\"}"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"pty-spawn","name":"terminal_open","arguments":"{\"type\":\"shell\",\"name\":\"main\"}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-spawn","name":"terminal_open","arguments":"{\"type\":\"shell\",\"name\":\"main\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"911213f8-acce-47be-a4f2-9d72ef55d83a"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[9,10,11,12,13],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-spawn","name":"terminal_open","arguments":"{\"type\":\"shell\",\"name\":\"main\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"36ae38c0-78e0-4302-9457-dd88fd8bde7c"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[9,10,11,12,13],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":1,"callId":"pty-spawn","name":"terminal_open","arguments":"{\"type\":\"shell\",\"name\":\"main\"}"}} -{"type":"tool/result","data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"pty-spawn"},"content":[{"type":"tool-result","toolCallId":"pty-spawn","content":[{"type":"text","text":"started terminal session pty-1 (main) [type: shell]\ndsh> "}],"isError":false}],"role":"user","id":"2da21b47-7fb3-444c-99a6-2c21743731ee"}},"sourceEventSeqs":[15],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":1,"message":{"source":{"kind":"tool","callId":"pty-spawn"},"content":[{"type":"tool-result","toolCallId":"pty-spawn","content":[{"type":"text","text":"started terminal session pty-1 (main) [type: shell]\ndsh> "}],"isError":false}],"role":"user","id":"f6583921-7e08-49f6-b5bd-99f6a3ced4b2"}},"sourceEventSeqs":[15],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":1}} {"type":"step/start","data":{"turn":1,"step":2}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -23,9 +23,9 @@ {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"pty-send","name":"terminal_send","arguments":"{\"sessionId\":\"pty-1\",\"text\":\"printf 'PTY_OK\\\\n'\"}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} {"type":"assistant/chunk","data":{"turn":1,"step":2,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":2,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-send","name":"terminal_send","arguments":"{\"sessionId\":\"pty-1\",\"text\":\"printf 'PTY_OK\\\\n'\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"cf12d7ee-322a-4057-97b0-98d828a96f1a"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[19,20,21,22,23],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":2,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-send","name":"terminal_send","arguments":"{\"sessionId\":\"pty-1\",\"text\":\"printf 'PTY_OK\\\\n'\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"000f1cda-e29a-42e1-9c7c-6e3803aa9316"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[19,20,21,22,23],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":2,"callId":"pty-send","name":"terminal_send","arguments":"{\"sessionId\":\"pty-1\",\"text\":\"printf 'PTY_OK\\\\n'\"}"}} -{"type":"tool/result","data":{"turn":1,"step":2,"message":{"source":{"kind":"tool","callId":"pty-send"},"content":[{"type":"tool-result","toolCallId":"pty-send","content":[{"type":"text","text":"K\ndsh> \n[wait: stdin_read]\n[session: running]\n[output truncated]"}],"isError":false}],"role":"user","id":"d1ffb2dc-6a33-4c9e-aeb9-b87bf6da8617"},"meta":{"viewport":"printf 'PTY_OK\\n'\nPTY_OK\ndsh> ","waitReason":"stdin_read","sessionStatus":{"kind":"running"},"truncated":false}},"sourceEventSeqs":[25],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":2,"message":{"source":{"kind":"tool","callId":"pty-send"},"content":[{"type":"tool-result","toolCallId":"pty-send","content":[{"type":"text","text":"K\ndsh> \n[wait: stdin_read]\n[session: running]\n[output truncated]"}],"isError":false}],"role":"user","id":"68ac57a2-530b-4479-b278-78b8c10657d5"},"meta":{"viewport":"printf 'PTY_OK\\n'\nPTY_OK\ndsh> ","waitReason":"stdin_read","sessionStatus":{"kind":"running"},"truncated":false}},"sourceEventSeqs":[25],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":2}} {"type":"step/start","data":{"turn":1,"step":3}} {"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -33,9 +33,9 @@ {"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"pty-read","name":"terminal_read","arguments":"{\"sessionId\":\"pty-1\",\"offset\":0,\"count\":20}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} {"type":"assistant/chunk","data":{"turn":1,"step":3,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":3,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-read","name":"terminal_read","arguments":"{\"sessionId\":\"pty-1\",\"offset\":0,\"count\":20}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"b9b99ae9-4685-4ee5-b951-fbe58f84c4e3"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[29,30,31,32,33],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":3,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-read","name":"terminal_read","arguments":"{\"sessionId\":\"pty-1\",\"offset\":0,\"count\":20}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"c67744f1-e874-4678-8ea4-45eb96d95692"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[29,30,31,32,33],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":3,"callId":"pty-read","name":"terminal_read","arguments":"{\"sessionId\":\"pty-1\",\"offset\":0,\"count\":20}"}} -{"type":"tool/result","data":{"turn":1,"step":3,"message":{"source":{"kind":"tool","callId":"pty-read"},"content":[{"type":"tool-result","toolCallId":"pty-read","content":[{"type":"text","text":"dsh> printf 'PTY_OK\\n'\nPTY_OK\ndsh> \n[lines: 0-3 of 3]"}],"isError":false}],"role":"user","id":"d0d7331d-0d54-457a-9f7b-beb22abd34e6"}},"sourceEventSeqs":[35],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":3,"message":{"source":{"kind":"tool","callId":"pty-read"},"content":[{"type":"tool-result","toolCallId":"pty-read","content":[{"type":"text","text":"dsh> printf 'PTY_OK\\n'\nPTY_OK\ndsh> \n[lines: 0-3 of 3]"}],"isError":false}],"role":"user","id":"e77a5157-0844-48d2-9a4d-4b242c5d8357"}},"sourceEventSeqs":[35],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":3}} {"type":"step/start","data":{"turn":1,"step":4}} {"type":"assistant/chunk","data":{"turn":1,"step":4,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -43,9 +43,9 @@ {"type":"assistant/chunk","data":{"turn":1,"step":4,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"pty-signal","name":"terminal_signal","arguments":"{\"sessionId\":\"pty-missing\",\"signal\":\"SIGINT\"}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":4,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} {"type":"assistant/chunk","data":{"turn":1,"step":4,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":4,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-signal","name":"terminal_signal","arguments":"{\"sessionId\":\"pty-missing\",\"signal\":\"SIGINT\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"9bb8ec3d-6e6f-44a2-8957-8f2d855f4834"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[39,40,41,42,43],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":4,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-signal","name":"terminal_signal","arguments":"{\"sessionId\":\"pty-missing\",\"signal\":\"SIGINT\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"cc545337-43b3-4fc6-8a17-74da2a49086a"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[39,40,41,42,43],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":4,"callId":"pty-signal","name":"terminal_signal","arguments":"{\"sessionId\":\"pty-missing\",\"signal\":\"SIGINT\"}"}} -{"type":"tool/result","data":{"turn":1,"step":4,"message":{"source":{"kind":"tool","callId":"pty-signal"},"content":[{"type":"tool-result","toolCallId":"pty-signal","content":[{"type":"text","text":"Error: unknown PTY session pty-missing"}],"isError":true}],"role":"user","id":"16fc1ca2-0eca-42d7-85d0-31794424c260"}},"sourceEventSeqs":[45],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":4,"message":{"source":{"kind":"tool","callId":"pty-signal"},"content":[{"type":"tool-result","toolCallId":"pty-signal","content":[{"type":"text","text":"Error: unknown PTY session pty-missing"}],"isError":true}],"role":"user","id":"3cde95e4-fcc3-422d-ae60-cc041c2cb223"}},"sourceEventSeqs":[45],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":4}} {"type":"step/start","data":{"turn":1,"step":5}} {"type":"assistant/chunk","data":{"turn":1,"step":5,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -53,9 +53,9 @@ {"type":"assistant/chunk","data":{"turn":1,"step":5,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"pty-kill","name":"terminal_close","arguments":"{\"sessionId\":\"pty-1\"}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":5,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} {"type":"assistant/chunk","data":{"turn":1,"step":5,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":5,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-kill","name":"terminal_close","arguments":"{\"sessionId\":\"pty-1\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"4ebdc957-0369-4bbb-a5a5-4d2ef8ac3493"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[49,50,51,52,53],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":5,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-kill","name":"terminal_close","arguments":"{\"sessionId\":\"pty-1\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"830096ca-82a9-4f44-b798-7d7040c35172"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[49,50,51,52,53],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":5,"callId":"pty-kill","name":"terminal_close","arguments":"{\"sessionId\":\"pty-1\"}"}} -{"type":"tool/result","data":{"turn":1,"step":5,"message":{"source":{"kind":"tool","callId":"pty-kill"},"content":[{"type":"tool-result","toolCallId":"pty-kill","content":[{"type":"text","text":"closed terminal session pty-1"}],"isError":false}],"role":"user","id":"55a3e2cc-dbc5-44bf-a824-e3bbc8568cd5"}},"sourceEventSeqs":[55],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":5,"message":{"source":{"kind":"tool","callId":"pty-kill"},"content":[{"type":"tool-result","toolCallId":"pty-kill","content":[{"type":"text","text":"closed terminal session pty-1"}],"isError":false}],"role":"user","id":"f43c7146-3cb9-4e91-96f8-aa77b26d09e2"}},"sourceEventSeqs":[55],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":5}} {"type":"step/start","data":{"turn":1,"step":6}} {"type":"assistant/chunk","data":{"turn":1,"step":6,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} @@ -63,9 +63,9 @@ {"type":"assistant/chunk","data":{"turn":1,"step":6,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"pty-list","name":"terminal_list","arguments":"{}"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":6,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} {"type":"assistant/chunk","data":{"turn":1,"step":6,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":1,"step":6,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-list","name":"terminal_list","arguments":"{}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"6f956d23-5437-4a75-93a9-3abacd378e07"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[59,60,61,62,63],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":6,"message":{"role":"assistant","content":[{"type":"tool-call","id":"pty-list","name":"terminal_list","arguments":"{}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"3070bc05-8897-4f63-8c9d-257e6b518b67"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[59,60,61,62,63],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":1,"step":6,"callId":"pty-list","name":"terminal_list","arguments":"{}"}} -{"type":"tool/result","data":{"turn":1,"step":6,"message":{"source":{"kind":"tool","callId":"pty-list"},"content":[{"type":"tool-result","toolCallId":"pty-list","content":[{"type":"text","text":"(no terminal sessions)"}],"isError":false}],"role":"user","id":"be12d914-6fe1-4c1e-8262-65b2aa9c20e5"}},"sourceEventSeqs":[65],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":1,"step":6,"message":{"source":{"kind":"tool","callId":"pty-list"},"content":[{"type":"tool-result","toolCallId":"pty-list","content":[{"type":"text","text":"(no terminal sessions)"}],"isError":false}],"role":"user","id":"bc65b492-2cef-4fe6-8476-910bdc21d4ed"}},"sourceEventSeqs":[65],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":6}} {"type":"step/start","data":{"turn":1,"step":7}} {"type":"assistant/chunk","data":{"turn":1,"step":7,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} @@ -73,6 +73,6 @@ {"type":"assistant/chunk","data":{"turn":1,"step":7,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"DONE"}}}} {"type":"assistant/chunk","data":{"turn":1,"step":7,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":3}}}} {"type":"assistant/chunk","data":{"turn":1,"step":7,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} -{"type":"assistant/message","data":{"turn":1,"step":7,"message":{"role":"assistant","content":[{"type":"text","text":"DONE"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"9cca9680-5795-47d8-8edc-f6d44bcaa1ef"},"usage":{"inputTokens":10,"outputTokens":3}},"sourceEventSeqs":[69,70,71,72,73],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":1,"step":7,"message":{"role":"assistant","content":[{"type":"text","text":"DONE"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"ce300b5f-bfbf-4d85-8801-dce12f260d36"},"usage":{"inputTokens":10,"outputTokens":3}},"sourceEventSeqs":[69,70,71,72,73],"surfaceOp":"append"} {"type":"step/end","data":{"turn":1,"step":7}} {"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} diff --git a/examples/headless-agent/tests/snapshots/subagent-settlement/child.expected.jsonl b/examples/headless-agent/tests/snapshots/subagent-settlement/child.expected.jsonl index f07a82ecbc..d2ac2d31c6 100644 --- a/examples/headless-agent/tests/snapshots/subagent-settlement/child.expected.jsonl +++ b/examples/headless-agent/tests/snapshots/subagent-settlement/child.expected.jsonl @@ -1,5 +1,5 @@ {"type":"session","version":0,"id":"{{sessionId}}","createdAt":0,"cwd":"{{cwd}}","parentSession":"{{sessionId}}","origin":"subagent","delegationDepth":1} -{"type":"subagent/descriptor","data":{"version":2,"mode":"continuable","provider":"spawn","label":"Return child result","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"continuable","provider":"spawn","label":"Return child result","agentProvider":"deepseek-official","agentModel":"deepseek-v4-flash"}} {"type":"session/end-seed","data":{}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly CHILD_RESULT and nothing else. Do not call report."}],"source":{"kind":"user"},"role":"user","id":"{{sessionId}}"}]}} {"type":"turn/start","data":{"turn":1}} diff --git a/examples/headless-agent/tests/subagent-inheritance-snapshots/parent-override/child.expected.jsonl b/examples/headless-agent/tests/subagent-inheritance-snapshots/parent-override/child.expected.jsonl index 013d23d9ee..286ae33dcc 100644 --- a/examples/headless-agent/tests/subagent-inheritance-snapshots/parent-override/child.expected.jsonl +++ b/examples/headless-agent/tests/subagent-inheritance-snapshots/parent-override/child.expected.jsonl @@ -3,12 +3,12 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Use the write tool exactly once with file_path set to exactly the relative path inherited.txt and content escaped. If the write is denied, reply with the single word CHILD_DENIED and the denial marker line; do not retry and do not request escalation. If it succeeds, reply CHILD_WROTE."}],"source":{"kind":"user"},"role":"user","id":"{{sessionId}}"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Delegated write probe"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Delegated write probe"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Use the write tool exactly once with file_path set to exactly the relative path inherited.txt and content escaped. If the write is denied, reply with the single word CHILD_DENIED and the denial marker line; do not retry and do not request escalation. If it succeeds, reply CHILD_WROTE."}],"source":{"kind":"user"},"role":"user","id":"{{sessionId}}"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: read-only. Any available operation enforced by the DSH file sandbox cannot modify files in the standing mode. Do not refuse a required modification from this policy alone: try an available tool normally and follow any denial and escalation guidance it returns.\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: read-only. Any available operation enforced by the DSH file sandbox cannot modify files in the standing mode. Do not refuse a required modification from this policy alone: try an available tool normally and follow any denial and escalation guidance it returns."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"{{sessionId}}"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Use the write tool exactly","messageSeqs":[6],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash","reasoningEffort":"low"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"tool-call-delta","index":0,"id":"child-write","name":"write","argumentsDelta":"{\"file_path\": \"inherited.txt\", \"content\": \"escaped\"}"}}} diff --git a/examples/headless-agent/tests/subagent-inheritance-snapshots/parent-override/parent.expected.jsonl b/examples/headless-agent/tests/subagent-inheritance-snapshots/parent-override/parent.expected.jsonl index 9efed81901..332cf1be07 100644 --- a/examples/headless-agent/tests/subagent-inheritance-snapshots/parent-override/parent.expected.jsonl +++ b/examples/headless-agent/tests/subagent-inheritance-snapshots/parent-override/parent.expected.jsonl @@ -2,6 +2,7 @@ {"type":"turn/start","data":{"turn":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Tighten this session to read-only."}],"source":{"kind":"user"},"role":"user","id":"{{sessionId}}"},"surfaceOp":"append"} {"type":"sandbox/mode","data":{"mode":"read-only"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash","reasoningEffort":"low"}},"reason":"initial"}} {"type":"turn/end","data":{"turn":1,"reason":{"kind":"completed"}}} {"type":"session/end-seed","data":{}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Delegate the write probe to a subagent."}],"source":{"kind":"user"},"role":"user","id":"{{sessionId}}"}]}} @@ -11,16 +12,16 @@ {"type":"user/message","data":{"content":[{"type":"text","text":"Delegate the write probe to a subagent."}],"source":{"kind":"user"},"role":"user","id":"{{sessionId}}"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: read-only. Any available operation enforced by the DSH file sandbox cannot modify files in the standing mode. Do not refuse a required modification from this policy alone: try an available tool normally and follow any denial and escalation guidance it returns."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: read-only. Any available operation enforced by the DSH file sandbox cannot modify files in the standing mode. Do not refuse a required modification from this policy alone: try an available tool normally and follow any denial and escalation guidance it returns."}]},"role":"user","id":"{{sessionId}}"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Tighten this session to read-only.","messageSeqs":[1],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash"},"system":"{{system}}","tools":"{{tools}}"},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"deepseek-v4-flash","reasoningEffort":"low"},"system":"{{system}}","tools":"{{tools}}"},"reason":"resume"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"deepseek-v4-flash"}} {"type":"assistant/chunk","data":{"turn":2,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"tool-call"}}} {"type":"assistant/chunk","data":{"turn":2,"step":1,"chunk":{"type":"tool-call-delta","index":0,"id":"delegate-write","name":"subagent","argumentsDelta":"{\"description\": \"Delegated write probe\", \"prompt\": \"Use the write tool exactly once with file_path set to exactly the relative path inherited.txt and content escaped. If the write is denied, reply with the single word CHILD_DENIED and the denial marker line; do not retry and do not request escalation. If it succeeds, reply CHILD_WROTE.\"}"}}} {"type":"assistant/chunk","data":{"turn":2,"step":1,"chunk":{"type":"block-end","index":0,"block":{"type":"tool-call","id":"delegate-write","name":"subagent","arguments":"{\"description\": \"Delegated write probe\", \"prompt\": \"Use the write tool exactly once with file_path set to exactly the relative path inherited.txt and content escaped. If the write is denied, reply with the single word CHILD_DENIED and the denial marker line; do not retry and do not request escalation. If it succeeds, reply CHILD_WROTE.\"}"}}}} {"type":"assistant/chunk","data":{"turn":2,"step":1,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} {"type":"assistant/chunk","data":{"turn":2,"step":1,"chunk":{"type":"finish","reason":{"kind":"tool-calls"}}}} -{"type":"assistant/message","data":{"turn":2,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"delegate-write","name":"subagent","arguments":"{\"description\": \"Delegated write probe\", \"prompt\": \"Use the write tool exactly once with file_path set to exactly the relative path inherited.txt and content escaped. If the write is denied, reply with the single word CHILD_DENIED and the denial marker line; do not retry and do not request escalation. If it succeeds, reply CHILD_WROTE.\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"{{sessionId}}"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[14,15,16,17,18],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":2,"step":1,"message":{"role":"assistant","content":[{"type":"tool-call","id":"delegate-write","name":"subagent","arguments":"{\"description\": \"Delegated write probe\", \"prompt\": \"Use the write tool exactly once with file_path set to exactly the relative path inherited.txt and content escaped. If the write is denied, reply with the single word CHILD_DENIED and the denial marker line; do not retry and do not request escalation. If it succeeds, reply CHILD_WROTE.\"}"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"{{sessionId}}"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[15,16,17,18,19],"surfaceOp":"append"} {"type":"tool/call","data":{"turn":2,"step":1,"callId":"delegate-write","name":"subagent","arguments":"{\"description\": \"Delegated write probe\", \"prompt\": \"Use the write tool exactly once with file_path set to exactly the relative path inherited.txt and content escaped. If the write is denied, reply with the single word CHILD_DENIED and the denial marker line; do not retry and do not request escalation. If it succeeds, reply CHILD_WROTE.\"}"}} -{"type":"tool/result","data":{"turn":2,"step":1,"message":{"source":{"kind":"tool","callId":"delegate-write"},"content":[{"type":"tool-result","toolCallId":"delegate-write","content":[{"type":"text","text":"CHILD_DENIED [sandbox: file access denied under read-only mode]"}],"isError":false}],"role":"user","id":"{{sessionId}}"}},"sourceEventSeqs":[20],"surfaceOp":"append"} +{"type":"tool/result","data":{"turn":2,"step":1,"message":{"source":{"kind":"tool","callId":"delegate-write"},"content":[{"type":"tool-result","toolCallId":"delegate-write","content":[{"type":"text","text":"CHILD_DENIED [sandbox: file access denied under read-only mode]"}],"isError":false}],"role":"user","id":"{{sessionId}}"}},"sourceEventSeqs":[21],"surfaceOp":"append"} {"type":"step/end","data":{"turn":2,"step":1}} {"type":"step/start","data":{"turn":2,"step":2}} {"type":"assistant/chunk","data":{"turn":2,"step":2,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} @@ -28,6 +29,6 @@ {"type":"assistant/chunk","data":{"turn":2,"step":2,"chunk":{"type":"block-end","index":0,"block":{"type":"text","text":"The delegated child was denied by the sandbox. PARENT_DONE"}}}} {"type":"assistant/chunk","data":{"turn":2,"step":2,"chunk":{"type":"usage","usage":{"inputTokens":10,"outputTokens":5}}}} {"type":"assistant/chunk","data":{"turn":2,"step":2,"chunk":{"type":"finish","reason":{"kind":"stop"}}}} -{"type":"assistant/message","data":{"turn":2,"step":2,"message":{"role":"assistant","content":[{"type":"text","text":"The delegated child was denied by the sandbox. PARENT_DONE"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"{{sessionId}}"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[24,25,26,27,28],"surfaceOp":"append"} +{"type":"assistant/message","data":{"turn":2,"step":2,"message":{"role":"assistant","content":[{"type":"text","text":"The delegated child was denied by the sandbox. PARENT_DONE"}],"source":{"kind":"model","provider":"deepseek-official","model":"deepseek-v4-flash"},"id":"{{sessionId}}"},"usage":{"inputTokens":10,"outputTokens":5}},"sourceEventSeqs":[25,26,27,28,29],"surfaceOp":"append"} {"type":"step/end","data":{"turn":2,"step":2}} {"type":"turn/end","data":{"turn":2,"reason":{"kind":"completed"}}} diff --git a/examples/headless-agent/tests/subagent-inheritance.snapshot.ts b/examples/headless-agent/tests/subagent-inheritance.snapshot.ts index 8ca9855ed4..7b68ab4481 100644 --- a/examples/headless-agent/tests/subagent-inheritance.snapshot.ts +++ b/examples/headless-agent/tests/subagent-inheritance.snapshot.ts @@ -9,7 +9,7 @@ import { fileURLToPath } from 'node:url' import { Context } from '@deepseek-ai/cordis' import { normalizeSessionSnapshot, type NormalizeContext } from '@deepseek-ai/dsh-acp-snapshot' import { LOADER_SMOKE_TEST_TIMEOUT_MS, runLoaderSmoke } from '@deepseek-ai/dsh-loader-smoke' -import { createUserMessage } from '@deepseek-ai/dsh-llm' +import { createUserMessage, ReasoningEffortId } from '@deepseek-ai/dsh-llm' import SessionStore, { SESSION_FORMAT_VERSION, SessionId, type SessionEvent, type SessionHeader } from '@deepseek-ai/dsh-session' import JsonlSessionPersistence from '@deepseek-ai/dsh-session-persistence-jsonl' import { describe, expect, it } from 'vitest' @@ -26,7 +26,7 @@ const sessionId = SessionId('subagent-inheritance-parent') const refreshing = process.env.DSH_SNAPSHOT === 'refresh' const task = 'Delegate the write probe to a subagent.' -/** Seed a completed parent turn with the only read-only fact in the app. */ +/** Seed a completed parent turn with its read-only policy and current LLM selection. */ async function seedReadOnlyParent(root: string, cwd: string): Promise { const ctx = new Context() await ctx.plugin(SessionStore) @@ -42,7 +42,22 @@ async function seedReadOnlyParent(root: string, cwd: string): Promise { { type: 'turn/start', seq: 0, time: 10, data: { turn: 1 } }, { type: 'user/message', seq: 1, time: 11, data: createUserMessage({ content: [{ type: 'text', text: 'Tighten this session to read-only.' }], source: { kind: 'user' } }), surfaceOp: 'append' }, { type: 'sandbox/mode', seq: 2, time: 12, data: { mode: 'read-only' } }, - { type: 'turn/end', seq: 3, time: 13, data: { turn: 1, reason: { kind: 'completed' } } }, + { + type: 'request/header', + seq: 3, + time: 13, + data: { + header: { + config: { + provider: 'deepseek-official', + model: 'deepseek-v4-flash', + reasoningEffort: ReasoningEffortId('low'), + }, + }, + reason: 'initial', + }, + }, + { type: 'turn/end', seq: 4, time: 14, data: { turn: 1, reason: { kind: 'completed' } } }, ] try { await ctx.sessionPersistence.create(meta) diff --git a/examples/python-sdk-agent/cordis.yml b/examples/python-sdk-agent/cordis.yml index 6a04f5f42d..eeececff10 100644 --- a/examples/python-sdk-agent/cordis.yml +++ b/examples/python-sdk-agent/cordis.yml @@ -69,6 +69,7 @@ config: provider: spawn toolName: subagent + enableModelSelection: true enableRunInBackground: false - id: tool-todo diff --git a/examples/python-sdk-agent/tests/keyless-smoke.e2e.ts b/examples/python-sdk-agent/tests/keyless-smoke.e2e.ts index 625fa2f16f..230ada0336 100644 --- a/examples/python-sdk-agent/tests/keyless-smoke.e2e.ts +++ b/examples/python-sdk-agent/tests/keyless-smoke.e2e.ts @@ -144,7 +144,9 @@ describe('Python SDK dsh profile keyless smoke', () => { }, }, }) + const tools = modelRequests[0]?.tools as { function?: { name?: string } }[] expect(modelRequests[0]?.max_tokens).toBe(1234) + expect(tools.map(tool => tool.function?.name)).toContain('list_subagent_models') child.stdin.write(`${JSON.stringify({ jsonrpc: '2.0', id: 3, method: 'shutdown' })}\n`) const shutdown = await waitForLine(lines, value => value.id === 3, () => stderr) diff --git a/examples/python-sdk-agent/tests/snapshots/subagent-spawn-in-process/notifications.expected.jsonl b/examples/python-sdk-agent/tests/snapshots/subagent-spawn-in-process/notifications.expected.jsonl index f00b85e15f..48d98ad6de 100644 --- a/examples/python-sdk-agent/tests/snapshots/subagent-spawn-in-process/notifications.expected.jsonl +++ b/examples/python-sdk-agent/tests/snapshots/subagent-spawn-in-process/notifications.expected.jsonl @@ -104,7 +104,7 @@ {"method":"session.status","params":{"sessionId":"{{sessionId}}","status":"running"}} {"method":"session.event","params":{"sessionId":"{{sessionId}}","event":{"type":"turn/start","seq":3,"time":0,"data":{"turn":1}}}} {"method":"session.event","params":{"sessionId":"{{sessionId}}","event":{"type":"agent/inbox/spliced","seq":4,"time":0,"data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}}}} -{"method":"session.event","params":{"sessionId":"{{sessionId}}","event":{"type":"subagent/descriptor","seq":5,"time":0,"data":{"version":2,"mode":"one-shot","provider":"spawn","label":"echo probe"}}}} +{"method":"session.event","params":{"sessionId":"{{sessionId}}","event":{"type":"subagent/descriptor","seq":5,"time":0,"data":{"version":3,"mode":"one-shot","provider":"spawn","label":"echo probe"}}}} {"method":"session.event","params":{"sessionId":"{{sessionId}}","event":{"type":"step/start","seq":6,"time":0,"data":{"turn":1,"step":1}}}} {"method":"session.event","params":{"sessionId":"{{sessionId}}","event":{"type":"user/message","seq":7,"time":0,"data":{"content":[{"type":"text","text":"Reply with exactly: child answer 42."}],"source":{"kind":"user"},"role":"user","id":"{{sessionId}}"},"surfaceOp":"append"}}} {"method":"session.event","params":{"sessionId":"{{sessionId}}","event":{"type":"user/message","seq":8,"time":0,"data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: workspace-write. Any available operation enforced by the DSH file sandbox may modify files under the session workspace: \"{{cwd}}\". Some platform temporary areas may also be writable.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: workspace-write. Any available operation enforced by the DSH file sandbox may modify files under the session workspace: \"{{cwd}}\". Some platform temporary areas may also be writable."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"{{sessionId}}"},"surfaceOp":"append"}}} diff --git a/examples/python-sdk-agent/tests/snapshots/subagent-spawn-in-process/session.1.jsonl b/examples/python-sdk-agent/tests/snapshots/subagent-spawn-in-process/session.1.jsonl index 52aa539f93..059bded5dc 100644 --- a/examples/python-sdk-agent/tests/snapshots/subagent-spawn-in-process/session.1.jsonl +++ b/examples/python-sdk-agent/tests/snapshots/subagent-spawn-in-process/session.1.jsonl @@ -4,7 +4,7 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly: child answer 42."}],"source":{"kind":"user"},"role":"user","id":"7ae1698c-db1d-4fca-8404-3a9dece9c1d0"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"echo probe"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"echo probe"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly: child answer 42."}],"source":{"kind":"user"},"role":"user","id":"7ae1698c-db1d-4fca-8404-3a9dece9c1d0"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: workspace-write. Any available operation enforced by the DSH file sandbox may modify files under the session workspace: \"{{cwd}}\". Some platform temporary areas may also be writable.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: workspace-write. Any available operation enforced by the DSH file sandbox may modify files under the session workspace: \"{{cwd}}\". Some platform temporary areas may also be writable."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"dc291267-28a7-40f4-adac-cd856dbe0bba"},"surfaceOp":"append"} diff --git a/packages/api/session-controller/tests/session-cold.host.spec.ts b/packages/api/session-controller/tests/session-cold.host.spec.ts index da2d77ece5..aa51d159f6 100644 --- a/packages/api/session-controller/tests/session-cold.host.spec.ts +++ b/packages/api/session-controller/tests/session-cold.host.spec.ts @@ -15,6 +15,7 @@ import { SessionHistoryController } from '@deepseek-ai/dsh-api-session-controlle import { TypertLookupFailure } from '@deepseek-ai/dsh-typert-protocol' import TypertRegistry from '@deepseek-ai/dsh-typert-registry' import { createUserMessage, MessageId } from '@deepseek-ai/dsh-llm' +import { snapshotSubagentDescriptor } from '@deepseek-ai/dsh-subagent' import type { Agent } from '@deepseek-ai/dsh-agent' import type { SessionEvent, SessionHeader, SessionId } from '@deepseek-ai/dsh-session' import type { SessionPromptRequest, SessionRequestId } from '../src/types.ts' @@ -448,7 +449,11 @@ describe('subagent ownership fence', () => { type: 'subagent/descriptor', seq: 2, time: 3, - data: { version: 2, mode: 'continuable', provider: 'spawn', label: 'child' }, + data: snapshotSubagentDescriptor({ + mode: 'continuable', + provider: 'spawn', + label: 'child', + }), }, { type: 'turn/end', seq: 3, time: 4, data: { turn: 1, reason: { kind: 'completed' } } }, ] as SessionEvent[] diff --git a/packages/bundle/base/cordis.patch.yml b/packages/bundle/base/cordis.patch.yml index 65da35f554..981e791fb4 100644 --- a/packages/bundle/base/cordis.patch.yml +++ b/packages/bundle/base/cordis.patch.yml @@ -327,12 +327,15 @@ config: provider: spawn toolName: subagent + enableModelSelection: true backgroundMode: continuable - # Fork stays one-shot: a continuable child's `report` tool and prompt - # section precede the inherited history a fork exists to reuse; one-shot - # fork children install neither, keeping the parent's request prefix. - # See .agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md. + # Fork omits model selection so provider/model stay equal to the parent and + # the inherited history remains eligible for KV Cache reuse. It stays one-shot + # because a continuable child's `report` tool and prompt section precede that + # history and invalidate the same prefix. + # See .agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md + # and .agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md. - id: tool-subagent-fork name: '@deepseek-ai/dsh-tool-subagent' config: diff --git a/packages/bundle/web-app/cordis.patch.yml b/packages/bundle/web-app/cordis.patch.yml index f22884cc77..18a0d3911a 100644 --- a/packages/bundle/web-app/cordis.patch.yml +++ b/packages/bundle/web-app/cordis.patch.yml @@ -41,6 +41,11 @@ # `dsh.client` rows are the browser roster the modules node half scans into # window.__DSH_BOOT__; the modules row is simultaneously a host row. - insert: + # Host-owned opt-in sampled when a new Web session receives its preset + # delegation tools. The Models page edits this settings namespace. + - id: subagent-model-selection-settings + name: '@deepseek-ai/dsh-tool-subagent/model-selection-settings' + - id: code-runtime name: '@deepseek-ai/dsh-code-runtime-worker-thread' diff --git a/packages/bundle/web-app/package.json b/packages/bundle/web-app/package.json index 2f26a9daed..68c43da265 100644 --- a/packages/bundle/web-app/package.json +++ b/packages/bundle/web-app/package.json @@ -45,6 +45,7 @@ }, "dependencies": { "@deepseek-ai/dsh-agent-presets": "workspace:^", + "@deepseek-ai/dsh-tool-subagent": "workspace:^", "@deepseek-ai/dsh-api-remotes": "workspace:^", "@deepseek-ai/dsh-app-boot": "workspace:^", "@deepseek-ai/dsh-client-connection": "workspace:^", diff --git a/packages/client/ui-settings-models/README.i18n.yaml b/packages/client/ui-settings-models/README.i18n.yaml index 0eb68f97f2..add13d5562 100644 --- a/packages/client/ui-settings-models/README.i18n.yaml +++ b/packages/client/ui-settings-models/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/client/ui-settings-models/README.md -README.md: cc430c91b9c4124fc70d0ec205b5870012dd5554 -README.zh.md: 62358adbc6f055697efe33e29e579f0aec64efc3 +README.md: 0daa9c5288f169138a5961e8fd6bee7818dba89b +README.zh.md: b2f80b845302d29728a0574b82f565c96acf59f2 diff --git a/packages/client/ui-settings-models/README.md b/packages/client/ui-settings-models/README.md index cc430c91b9..0daa9c5288 100644 --- a/packages/client/ui-settings-models/README.md +++ b/packages/client/ui-settings-models/README.md @@ -4,6 +4,8 @@ English | [中文](README.zh.md) Models settings and product-onboarding plugin. The same client Cordis plugin registers the Models page plus two ordered first-run dialogs: a versioned internal-testing notice and the conditional official-DeepSeek credential step. Both steps share one modal wrapper and remain sequenced by `settings.onboarding`. The Models plane joins three wire domains into one shared snapshot — `llm.providers` (the configurable-provider directory with each route's live/dormant state), `settings.describe` (serialized schemas, layered redacted values, secret slots), and `credentials.describe` (value-free configured/source/writable badges) — and renders provider rows with one editor card at a time, without presenting route liveness as provider status. +When the Host advertises the `subagent-model-selection` settings namespace, Models also renders a localized switch above the provider rows. It defaults off and writes only `{ enabled }` through `settings.update` with the namespace revision. The Host samples it while composing a new top-level Session; changing it does not reconfigure running Sessions, while child Sessions inherit their parent's recorded decision. + Rows are the *configured* providers (their profile resolves in the owning namespace); a whole-section provider whose key is not configured anywhere renders as its open setup card instead of a row, but only in the first-run posture — while no provider is registered with the credential its profile names — and only until the user closes that card, after which it is an ordinary row carrying the missing-key dot. Each card kind owns its own open state, so closing one never discards a draft in another. The add flow is a card carrying the dormant-directory provider select — a bare-mounted `llm-pi-ai` offers its whole installed catalog before any route exists. The pi-ai card additionally edits that route's **model list** and can ask the provider what it serves. A row labels API-key state with a green solid dot only when a referenced credential is confirmed configured, and with a red solid dot only when a named reference is confirmed missing; reference-free provider-native authentication and unavailable credential enrichment remain unmarked. The editor is a hand-written card per adapter family: the primary field is a single **API key** input — the page never asks for an environment-variable name; a typed key stores **write-only** through `credentials.set` under the profile's reference, deriving `_API_KEY` when the profile has none, and the pi-ai profile records that derivation as `apiKeyEnv`, so `settings.yaml` never carries a key value. Leaving a new pi-ai provider's key blank saves a reference-free profile and therefore preserves provider-native authentication such as the Bedrock credential chain or Vertex ADC. A successful Apply emits a local accessible status message without echoing secret material. The collapsed 自定义设置 fold carries the curated extras — `baseURL` for both families (the deepseek placeholder shows the public endpoint), each adapter's model catalog, and the **display name** and **API protocol** of a pi-ai route the adapter does not ship. Those two are what a hand-declared route names for itself: the create card asks for both because nothing can default them, so the editor reaches both rather than leaving them to `settings.yaml`. Clearing the name unsets it and the route falls back to its id, which is what the placeholder shows; the protocol has no such fallback. A catalog route gets neither — it defaults its name from its catalog entry, and its models each carry their own protocol, so a route-level one could only override every one of them. The Provider ID stays fixed: it is the settings key, the name every other namespace and every logged session references, and the stem of a credential reference the page cannot read back to move. Reasoning effort is deliberately NOT among them: it is a per-model capability and the models under one provider disagree about which levels they accept, so a provider-scoped control could only be set to a value some of them reject — which would hide even the models that support the level. The composer's model picker offers each model its own levels, and a switch there records provider, model, and effort together as the default for the next session. The profile field stays in `settings.yaml` for a deployment that knows its route. Each DeepSeek row edits `id`, optional display `name`, and optional `contextWindow`/`maxTokens`; existing fields outside that curated set survive edits, while every other profile field stays owned by `settings.yaml`. A row is deletable only when the user layer alone carries it (removal restores the composition base), and its localized confirmation dialog names the provider in the title, description, and final action. A row is tagged **Custom** when the directory entry says the owning adapter ships nothing under that key. The tag follows that answer alone: having a stored profile does not make a route custom — narrowing a shipped provider's models stores one too — and an adapter that reports nothing leaves its rows untagged rather than being read as shipped. The notice step owns its exact copy in `src/client/locales.ts` and its acknowledgement version in `src/onboarding-copy.ts`. On loopback it compares and writes `ui-onboarding.welcomeNoticeVersion` through the existing settings API; only an explicit Continue records the current version. A non-loopback browser cannot use that Host-only namespace, so acknowledgement is process-local and the notice returns after reload. diff --git a/packages/client/ui-settings-models/README.zh.md b/packages/client/ui-settings-models/README.zh.md index 62358adbc6..b2f80b8453 100644 --- a/packages/client/ui-settings-models/README.zh.md +++ b/packages/client/ui-settings-models/README.zh.md @@ -4,6 +4,8 @@ 模型设置与产品引导插件。同一个 client Cordis 插件会注册 Models 页面和两个有序的首次使用弹窗：版本化内测声明，以及按条件显示的 DeepSeek 官方凭据步骤。两个步骤共用同一套弹窗组件，并继续由 `settings.onboarding` 排序。Models 平面把三个协议领域汇聚为一个共享快照：`llm.providers`（可配置提供方目录，含每条路由的存活／休眠状态）、`settings.describe`（序列化 schema、分层脱敏值、secret slot）与 `credentials.describe`（不含值的 configured/source/writable 徽标）；页面据此渲染提供方行，一次只展开一张编辑卡片，且不把路由存活状态呈现为提供方状态。 +Host 公布 `subagent-model-selection` settings namespace 时，Models 还会在提供方行上方渲染本地化开关。它默认关闭，只通过 `settings.update` 携带 namespace revision 写入 `{ enabled }`。Host 会在组合新的顶层 Session 时读取它；修改设置不会重新配置运行中的 Session，而子 Session 会继承父级已记录的决定。 + 行是*已配置*的提供方（其 profile 在所属 namespace 中解析得出）；其配置键未在任何位置配置的整分节提供方会渲染为其展开的设置卡片而非一行，但仅限首次运行姿态——即尚无任何提供方已注册且备齐其 profile 所指名的凭据——且仅持续到用户关闭该卡片为止，此后它就是一行带缺失密钥点的普通行。每一类卡片各自持有自己的展开状态，因此关掉其中一张绝不会丢弃另一张里的草稿。「新增」流程则是一张承载休眠目录提供方选择框的卡片——裸挂载的 `llm-pi-ai` 在任何路由存在之前就能提供其完整的已安装 catalog。pi-ai 卡片还会编辑该路由的**模型列表**，并可查询提供方所提供的模型。只有确认引用的凭据已配置时，行才会以绿色实心点标示 API 密钥状态；只有确认具名引用缺失时，才会以红色实心点标示。无引用的提供方原生认证以及无法取得凭据补充信息时都不显示状态点。编辑器是每个适配器家族各一张的手写卡片：主字段是单独一个 **API 密钥**输入框——页面从不询问环境变量名；键入的密钥经 `credentials.set` 以**只写**方式存入 profile 的引用之下，profile 没有引用时便派生 `_API_KEY`，pi-ai profile 会把这次派生记录为 `apiKeyEnv`，因此 `settings.yaml` 从不携带密钥值。为新的 pi-ai 提供方留空密钥会保存一个不带引用的 profile，因此能保留提供方原生认证，例如 Bedrock 凭据链或 Vertex ADC。「应用」成功后会发出本地无障碍状态消息，且绝不回显任何机密内容。收起的「自定义设置」折叠区承载精选的额外字段——两个家族都有 `baseURL`（deepseek 的占位符显示公共端点）、各适配器自己的模型目录，以及适配器未提供的那类 pi-ai 路由的**显示名称**与 **API 协议**。这两个字段是手工声明路由为自己命名的东西：创建卡片之所以索要它们，正因为没有东西能为它们兜底，因此编辑器也够得着这两个，而不是把它们留给 `settings.yaml`。清空名称即取消设置，路由退回自己的 id——占位符显示的就是它；协议没有这样的兜底。内置目录路由两个都不给：它的名称由目录条目兜底，它的每个模型各自带着自己的协议，路由级协议只可能把它们全部覆盖掉。Provider ID 保持固定：它是 settings 的键、是其他每个 namespace 与每一条已记录会话引用的名字，也是页面读不回、因而搬不走的凭据引用词干。推理等级刻意**不在**其中：它是按模型的能力，而同一提供方下各模型接受的档位并不一致，因此提供方级的控件只可能被设成其中一些模型会拒绝的值——那会连支持该档位的模型也一并隐藏。输入框的模型选择器为每个模型提供它自己的档位，在那里切换会把提供方、模型、推理等级一并记为下一个会话的默认值。profile 字段仍留在 `settings.yaml`，供清楚自己路由的部署使用。每条 DeepSeek 模型行可编辑 `id`、可选的显示名称 `name` 与可选的 `contextWindow`/`maxTokens`；精选集合以外的现有字段会在编辑后保留，其余每个 profile 字段仍归 `settings.yaml` 所有。只有当某行仅由用户层承载时它才可删除（删除会还原组合 base），其本地化确认对话框会在标题、说明和最终操作中点名该提供方。当目录条目表明拥有该路由的适配器在这个键下什么都没有时，该行会带上 **自定义** 标签。标签只跟随这个答案：存了 profile 并不使一条路由成为自定义——收窄一个内置提供方的模型同样会存下 profile——而什么都不回答的适配器，其路由保持无标签，不会被当成内置。 声明步骤在 `src/client/locales.ts` 中持有完整文案，并在 `src/onboarding-copy.ts` 中持有确认版本。回环访问会通过既有 settings API 比较并写入 `ui-onboarding.welcomeNoticeVersion`；只有明确点击「继续」才会记录当前版本。非回环浏览器无法使用这项仅限 Host 的 namespace，因此确认仅在当前进程有效，重载后声明会再次出现。 diff --git a/packages/client/ui-settings-models/src/client/ModelsSection.module.css b/packages/client/ui-settings-models/src/client/ModelsSection.module.css index 3719535a3a..1c767386dd 100644 --- a/packages/client/ui-settings-models/src/client/ModelsSection.module.css +++ b/packages/client/ui-settings-models/src/client/ModelsSection.module.css @@ -40,6 +40,87 @@ color: var(--dsw-alias-state-success-primary); } +.preferenceCard { + display: grid; + grid-template-columns: minmax(0, 1fr) auto; + align-items: center; + gap: 8px 16px; + margin-top: 4px; + padding: 14px; + border: 1px solid var(--dsw-alias-border-l2); + border-radius: 12px; +} + +.preferenceCopy { + min-width: 0; +} + +.preferenceTitle { + margin: 0; + font-size: 14px; + line-height: 22px; + font-weight: 500; + color: var(--dsw-alias-label-primary); +} + +.preferenceDescription { + margin: 2px 0 0; + font-size: 12px; + line-height: 18px; + color: var(--dsw-alias-label-tertiary); +} + +.switch { + box-sizing: border-box; + position: relative; + width: 36px; + height: 20px; + padding: 2px; + border: 0; + border-radius: 10px; + background: var(--dsw-alias-border-l3); + cursor: pointer; +} + +.switchOn { + background: var(--dsw-alias-brand-primary); +} + +.switch:disabled { + cursor: default; + opacity: 0.5; +} + +.switch:focus-visible { + outline: none; + box-shadow: 0 0 0 2px var(--dsw-alias-border-l3); +} + +.switchThumb { + display: block; + width: 16px; + height: 16px; + border-radius: 50%; + background: var(--dsw-alias-label-primary-foreground); + transition: transform 120ms ease; +} + +.switchOn .switchThumb { + transform: translateX(16px); +} + +.preferenceStatus, +.preferenceCard > .error { + grid-column: 1 / -1; +} + +.preferenceStatus { + margin: 0; + font-size: 12px; + line-height: 18px; + color: var(--dsw-alias-state-success-primary); +} + .rows { list-style: none; /* Extra air between the title/intro block and the first provider card. */ @@ -623,7 +704,8 @@ select.input { } @media (prefers-reduced-motion: reduce) { - .customizedSummary::before { + .customizedSummary::before, + .switchThumb { transition: none; } } diff --git a/packages/client/ui-settings-models/src/client/ModelsSection.tsx b/packages/client/ui-settings-models/src/client/ModelsSection.tsx index 7f178564d6..9501185e87 100644 --- a/packages/client/ui-settings-models/src/client/ModelsSection.tsx +++ b/packages/client/ui-settings-models/src/client/ModelsSection.tsx @@ -22,6 +22,7 @@ import { deriveKeyRef, messageOf, protocolChoices, providerUsable } from './stor import type { ModelsSettingsStore, ProviderRow } from './store.ts' import type { SettingsSchemaOperations } from './schema-operations.ts' import { ProviderEditor, type ProviderEditorProps } from './ProviderEditor.tsx' +import { SubagentModelSelectionCard } from './SubagentModelSelectionCard.tsx' import type { en } from './locales.ts' import styles from './ModelsSection.module.css' @@ -278,12 +279,24 @@ function Loaded({ injected }: { injected: ModelsSectionFace }): ReactNode { // one whose schema names the protocols one may speak; without it mounted // there is nothing to declare and the entry point stays disabled. const protocols = protocolChoices(state.namespaces.get('llm-pi-ai'), schema) + const subagentModelSelection = state.namespaces.get('subagent-model-selection') return (

{t('title')}

{t('intro')}

{!state.writable && state.status === 'ready' ?

{t('readOnly')}

: null} + {subagentModelSelection === undefined + ? null + : ( + + )} {savedIdentity === undefined ? null : ( diff --git a/packages/client/ui-settings-models/src/client/SubagentModelSelectionCard.tsx b/packages/client/ui-settings-models/src/client/SubagentModelSelectionCard.tsx new file mode 100644 index 0000000000..3b62f28e60 --- /dev/null +++ b/packages/client/ui-settings-models/src/client/SubagentModelSelectionCard.tsx @@ -0,0 +1,87 @@ +/** User control for model-selectable subagent delegation in new sessions. */ + +import { useState } from 'react' +import type { ReactNode } from 'react' +import type { IApiClient, SettingsNamespaceView } from '@deepseek-ai/dsh-api-remotes/client' +import type { ModelsSettingsStore } from './store.ts' +import type { en } from './locales.ts' +import { messageOf } from './store.ts' +import styles from './ModelsSection.module.css' + +/** Props for the Host-owned subagent model-selection preference. */ +export interface SubagentModelSelectionCardProps { + /** Current redacted namespace view. */ + namespace: SettingsNamespaceView + /** Whether the settings provider accepts writes. */ + writable: boolean + /** Settings wire face. */ + api: Pick + /** Models page controller to refresh after a commit. */ + controller: ModelsSettingsStore + /** Localized Models copy. */ + t: (key: keyof typeof en) => string +} + +/** Read the schema-validated resolved boolean from a namespace view. */ +function enabledOf(namespace: SettingsNamespaceView): boolean { + if (typeof namespace.value !== 'object' || namespace.value === null) return false + return (namespace.value as { enabled?: unknown }).enabled === true +} + +/** Render and persist the default-off new-session preference. */ +export function SubagentModelSelectionCard({ + namespace, + writable, + api, + controller, + t, +}: SubagentModelSelectionCardProps): ReactNode { + const [saving, setSaving] = useState(false) + const [saved, setSaved] = useState(false) + const [error, setError] = useState(undefined) + const enabled = enabledOf(namespace) + + const toggle = (): void => { + setSaving(true) + setSaved(false) + setError(undefined) + void api.settings.update({ + ns: namespace.ns, + patch: { enabled: !enabled }, + expectedRevision: namespace.revision, + }).then(async (response) => { + if (!response.result.ok) throw new Error(response.result.error.message) + controller.acceptNamespace(response.result.value) + await controller.load() + setSaved(true) + }).catch((reason: unknown) => { + setError(messageOf(reason)) + }).finally(() => { setSaving(false) }) + } + + return ( +
+
+

+ {t('subagentModelSelectionTitle')} +

+

{t('subagentModelSelectionDescription')}

+
+ + {saved + ?

{t('subagentModelSelectionSaved')}

+ : null} + {error === undefined ? null :

{error}

} +
+ ) +} diff --git a/packages/client/ui-settings-models/src/client/locales.ts b/packages/client/ui-settings-models/src/client/locales.ts index f1b0718ba5..176e33fe5e 100644 --- a/packages/client/ui-settings-models/src/client/locales.ts +++ b/packages/client/ui-settings-models/src/client/locales.ts @@ -5,6 +5,10 @@ export const en = { nav: 'Models', title: 'Models', intro: 'Enter your API keys to use models from the following providers.', + subagentModelSelectionTitle: 'Subagent model selection', + subagentModelSelectionDescription: 'Allow new sessions to choose a provider, model, and reasoning effort for subagents. Running sessions do not change.', + subagentModelSelectionToggle: 'Allow subagents to choose models', + subagentModelSelectionSaved: 'Saved. New sessions use this setting.', edit: 'Edit', editProvider: 'Edit {provider}', remove: 'Delete', @@ -109,6 +113,10 @@ export const zh: { [Key in keyof typeof en]: string } = { nav: '模型', title: '模型', intro: '填入各提供方的 API 密钥即可使用其模型。', + subagentModelSelectionTitle: 'Subagent 自选模型', + subagentModelSelectionDescription: '允许新会话为 subagent 选择提供方、模型和推理强度。运行中的会话不会改变。', + subagentModelSelectionToggle: '允许 subagent 自选模型', + subagentModelSelectionSaved: '已保存，新会话将使用此设置。', edit: '编辑', editProvider: '编辑 {provider}', remove: '删除', diff --git a/packages/client/ui-settings-models/src/client/store.ts b/packages/client/ui-settings-models/src/client/store.ts index 349798acdd..2b17c9fc4b 100644 --- a/packages/client/ui-settings-models/src/client/store.ts +++ b/packages/client/ui-settings-models/src/client/store.ts @@ -124,6 +124,15 @@ export class ModelsSettingsStore { private readonly describeFace: SettingsDescribeFace, ) {} + /** + * Fold one successful settings write into the shared mirror before rejoining + * this page's rows. + * @param view - namespace view returned by the settings wire method. + */ + acceptNamespace(view: SettingsNamespaceView): void { + this.describeFace.acceptView(view) + } + /** * Refresh the whole page snapshot: the provider directory and the mirror's * settings answer in parallel, then one batched credential describe over diff --git a/packages/client/ui-settings-models/tests/components.client.spec.tsx b/packages/client/ui-settings-models/tests/components.client.spec.tsx index 8a3b8fce0a..5d0cbe8fed 100644 --- a/packages/client/ui-settings-models/tests/components.client.spec.tsx +++ b/packages/client/ui-settings-models/tests/components.client.spec.tsx @@ -8,6 +8,7 @@ import type { RpcResponse, SettingsNamespaceView } from '@deepseek-ai/dsh-api-re import { ModelsSection, needsSetup, providerCopy, providerTargetLabel, removeProviderProfile, } from '../src/client/ModelsSection.tsx' +import { SubagentModelSelectionCard } from '../src/client/SubagentModelSelectionCard.tsx' import type { ModelsSectionInjected, ModelsSectionProps } from '../src/client/ModelsSection.tsx' import { pathOps } from '../src/client/ProviderEditor.tsx' import { @@ -122,6 +123,14 @@ function wireNamespaces(): SettingsNamespaceView[] { secrets: [], revision: 0, }, + { + ns: 'subagent-model-selection', + schema: JSON.parse(JSON.stringify(Schema.object({ enabled: Schema.boolean().default(false) }).toJSON())) as unknown, + value: { enabled: false }, + applies: 'live', + secrets: [], + revision: 4, + }, ] } @@ -143,9 +152,10 @@ function scriptedFace(overrides: { set?: ReturnType unset?: ReturnType } = {}) { - const update = overrides.update ?? vi.fn(() => Promise.resolve(ok(wireNamespaces()[2]))) - const replace = overrides.replace ?? vi.fn(() => Promise.resolve(ok(wireNamespaces()[2]))) - const mutate = overrides.mutate ?? vi.fn(() => Promise.resolve(ok(wireNamespaces()[2]))) + const providerNamespace = wireNamespaces().find(view => view.ns === 'llm-pi-ai')! + const update = overrides.update ?? vi.fn(() => Promise.resolve(ok(providerNamespace))) + const replace = overrides.replace ?? vi.fn(() => Promise.resolve(ok(providerNamespace))) + const mutate = overrides.mutate ?? vi.fn(() => Promise.resolve(ok(providerNamespace))) const set = overrides.set ?? vi.fn(() => Promise.resolve(ok({}))) const unset = overrides.unset ?? vi.fn(() => Promise.resolve(ok({}))) const face = { @@ -236,6 +246,71 @@ describe('ModelsSection', () => { expect(document.body.textContent).toBe('') }) + it('persists the default-off subagent model-selection switch for new sessions', async () => { + const enabledNamespace: SettingsNamespaceView = { + ...wireNamespaces().find(view => view.ns === 'subagent-model-selection')!, + value: { enabled: true }, + user: { enabled: true }, + revision: 5, + } + const update = vi.fn(() => Promise.resolve(ok(enabledNamespace))) + await mountSection({ update }) + + const toggle = screen.getByRole('switch', { name: en.subagentModelSelectionToggle }) + expect(toggle.getAttribute('aria-checked')).toBe('false') + fireEvent.click(toggle) + + await waitFor(() => { expect(toggle.getAttribute('aria-checked')).toBe('true') }) + expect(update).toHaveBeenCalledWith({ + ns: 'subagent-model-selection', + patch: { enabled: true }, + expectedRevision: 4, + }) + expect(screen.getByRole('status').textContent).toBe(en.subagentModelSelectionSaved) + }) + + it('reports rejected subagent model-selection updates and permits a retry', async () => { + const update = vi.fn() + .mockResolvedValueOnce(fail('revision changed')) + .mockResolvedValueOnce(ok({ + ...wireNamespaces().find(view => view.ns === 'subagent-model-selection')!, + value: { enabled: true }, + revision: 5, + })) + await mountSection({ update }) + + const toggle = screen.getByRole('switch', { name: en.subagentModelSelectionToggle }) + fireEvent.click(toggle) + expect((await screen.findByRole('alert')).textContent).toBe('revision changed') + + fireEvent.click(toggle) + await waitFor(() => { expect(toggle.getAttribute('aria-checked')).toBe('true') }) + expect(screen.queryByRole('alert')).toBeNull() + }) + + it('keeps malformed and read-only subagent preferences off', () => { + const namespace = { + ...wireNamespaces().find(view => view.ns === 'subagent-model-selection')!, + value: null, + } as unknown as SettingsNamespaceView + const update = vi.fn() + render( + , + ) + + const toggle = screen.getByRole('switch', { name: en.subagentModelSelectionToggle }) + expect(toggle.getAttribute('aria-checked')).toBe('false') + expect((toggle as HTMLButtonElement).disabled).toBe(true) + fireEvent.click(toggle) + expect(update).not.toHaveBeenCalled() + }) + it('renders the unkeyed whole-section provider as an open setup card in the first-run posture', async () => { await mountFirstRun() // Nothing is reachable yet, and DeepSeek has no configured credential and diff --git a/packages/core/agent-loop/README.i18n.yaml b/packages/core/agent-loop/README.i18n.yaml index 2a5445de6b..88d4e6da7d 100644 --- a/packages/core/agent-loop/README.i18n.yaml +++ b/packages/core/agent-loop/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/core/agent-loop/README.md -README.md: 907bbe75fd687388be044916f4c4509e1c552177 -README.zh.md: 76a73155a3d9f811931cf0f86c227e6ab3dc1c32 +README.md: 1b233ae1203171930ef5b58de93ec67381ec4918 +README.zh.md: 81af654072f23c5280e2e14bc891972b5e1f37d5 diff --git a/packages/core/agent-loop/README.md b/packages/core/agent-loop/README.md index 907bbe75fd..1b233ae120 100644 --- a/packages/core/agent-loop/README.md +++ b/packages/core/agent-loop/README.md @@ -42,6 +42,7 @@ interface Config { id: string // required provider?: string model?: string + reasoningEffort?: string // non-empty initial reasoning effort maxTokens?: number // positive per-request output-token cap resumeSessionId?: string // load this persisted session instead of creating one cwd?: string // optional workspace cwd for the fresh session @@ -49,7 +50,7 @@ interface Config { } ``` -Configured agents start automatically. A model call requires both `provider` and `model`; `agent/request` may supply a missing pair before dispatch. An optional positive `maxTokens` seeds each conversation request's output cap and is logged in its request header. `maxParallelToolCalls` bounds every agent's rolling pool for parallel-safe calls and defaults to `10`; it is also the whole of the `agent-loop` Settings section, so a user layer over this entry caps the next tool group without a restart, and a value that is not a positive integer is refused at the write rather than at that group. `agents` is deliberately absent from that section — it is consumed once when the service starts, so a stored change could only look like it had an effect. `cwd` applies only to fresh sessions, while `resumeSessionId` retains persisted metadata. Configured agents use the deployment persona, and programmatic setup can shadow it per agent. This plugin supplies the per-agent `provider`, `model`, and `cwd` prompt variables; harness identity and deployment persona belong to `dsh-system-prompt`. +Configured agents start automatically. A model call requires both `provider` and `model`; `agent/request` may supply a missing pair before dispatch. An optional non-empty `reasoningEffort` seeds the request's reasoning setting; `agent/request` may override it, and adapter resolution validates the effective value recorded in the request header. An optional positive `maxTokens` seeds each conversation request's output cap and is logged in its request header. `maxParallelToolCalls` bounds every agent's rolling pool for parallel-safe calls and defaults to `10`; it is also the whole of the `agent-loop` Settings section, so a user layer over this entry caps the next tool group without a restart, and a value that is not a positive integer is refused at the write rather than at that group. `agents` is deliberately absent from that section — it is consumed once when the service starts, so a stored change could only look like it had an effect. `cwd` applies only to fresh sessions, while `resumeSessionId` retains persisted metadata. Configured agents use the deployment persona, and programmatic setup can shadow it per agent. This plugin supplies the per-agent `provider`, `model`, and `cwd` prompt variables; harness identity and deployment persona belong to `dsh-system-prompt`. ### Internal concrete driver diff --git a/packages/core/agent-loop/README.zh.md b/packages/core/agent-loop/README.zh.md index 76a73155a3..81af654072 100644 --- a/packages/core/agent-loop/README.zh.md +++ b/packages/core/agent-loop/README.zh.md @@ -42,6 +42,7 @@ interface Config { id: string // required provider?: string model?: string + reasoningEffort?: string // non-empty initial reasoning effort maxTokens?: number // positive per-request output-token cap resumeSessionId?: string // load this persisted session instead of creating one cwd?: string // optional workspace cwd for the fresh session @@ -49,7 +50,7 @@ interface Config { } ``` -通过配置创建的 agent 会自动启动。模型调用同时需要 `provider` 和 `model`；`agent/request` 可以在分发前补齐缺失的这一对值。可选的正数 `maxTokens` 会为每次对话请求提供初始输出上限，并记录在请求 header 中。`maxParallelToolCalls` 限制每个 agent 针对并行安全调用使用的滚动池，默认值为 `10`；它同时也是 `agent-loop` Settings 段的全部内容，因此叠加在该条目之上的用户层无需重启即可限制下一组工具调用，而非正整数的值会在写入时被拒绝，而不是到那一组时才失败。`agents` 刻意不在该段中——它在服务启动时被消费一次，所以存储的改动只会看起来生效。`cwd` 仅应用于全新会话，而 `resumeSessionId` 保留持久化元数据。通过配置创建的 agent 使用部署 persona；编程式 setup 可以按 agent 遮蔽它。该插件为每个 agent 提供 `provider`、`model` 和 `cwd` 提示词变量；harness 身份与部署 persona 属于 `dsh-system-prompt`。 +通过配置创建的 agent 会自动启动。模型调用同时需要 `provider` 和 `model`；`agent/request` 可以在分发前补齐缺失的这一对值。可选的非空 `reasoningEffort` 会提供请求的初始推理强度；`agent/request` 可以覆盖它，适配器解析会校验记录在请求 header 中的最终值。可选的正数 `maxTokens` 会为每次对话请求提供初始输出上限，并记录在请求 header 中。`maxParallelToolCalls` 限制每个 agent 针对并行安全调用使用的滚动池，默认值为 `10`；它同时也是 `agent-loop` Settings 段的全部内容，因此叠加在该条目之上的用户层无需重启即可限制下一组工具调用，而非正整数的值会在写入时被拒绝，而不是到那一组时才失败。`agents` 刻意不在该段中——它在服务启动时被消费一次，所以存储的改动只会看起来生效。`cwd` 仅应用于全新会话，而 `resumeSessionId` 保留持久化元数据。通过配置创建的 agent 使用部署 persona；编程式 setup 可以按 agent 遮蔽它。该插件为每个 agent 提供 `provider`、`model` 和 `cwd` 提示词变量；harness 身份与部署 persona 属于 `dsh-system-prompt`。 ### 包内部具体驱动器 diff --git a/packages/core/agent-loop/src/agent.ts b/packages/core/agent-loop/src/agent.ts index 3ef1ec7aa4..6bf7517903 100644 --- a/packages/core/agent-loop/src/agent.ts +++ b/packages/core/agent-loop/src/agent.ts @@ -438,11 +438,12 @@ export class ReactLoopAgent implements Agent { const persistedHeader = session.requestHeader() const persistedConfig = persistedHeader?.config const route = { provider: this.options.provider ?? '', model: this.options.model ?? '' } - const reasoningEffort = persistedConfig?.provider === route.provider + const persistedReasoningEffort = persistedConfig?.provider === route.provider && persistedConfig.model === route.model && persistedHeader?.adapterDefaults?.reasoningEffort !== true ? persistedConfig.reasoningEffort : undefined + const reasoningEffort = this.options.reasoningEffort ?? persistedReasoningEffort const maxTokens = this.options.maxTokens const seedConfig = deepFreeze(structuredClone( this.requestHeaderLogged diff --git a/packages/core/agent-loop/src/index.ts b/packages/core/agent-loop/src/index.ts index 371154a7c9..38ee7dfef5 100644 --- a/packages/core/agent-loop/src/index.ts +++ b/packages/core/agent-loop/src/index.ts @@ -19,7 +19,7 @@ import type { ResumeAgentOptions, SessionStartSource, } from '@deepseek-ai/dsh-agent' -import { errorChain } from '@deepseek-ai/dsh-llm' +import { errorChain, ReasoningEffortId } from '@deepseek-ai/dsh-llm' import { installSettingsSection, settingsNamespace } from '@deepseek-ai/dsh-settings' import { SessionId, SessionPreparation } from '@deepseek-ai/dsh-session' import type { Session, SessionHeader } from '@deepseek-ai/dsh-session' @@ -304,6 +304,7 @@ export class AgentLoop extends Service implements AgentFactory { sessionId: z.string().min(1), provider: z.string(), model: z.string(), + reasoningEffort: z.string().min(1) as z>, maxTokens: z.number().step(1).min(1).max(Number.MAX_SAFE_INTEGER), cwd: z.string(), resumeSessionId: z.string(), diff --git a/packages/core/agent-loop/tests/loop.spec.ts b/packages/core/agent-loop/tests/loop.spec.ts index 2105b86f42..4082b43452 100644 --- a/packages/core/agent-loop/tests/loop.spec.ts +++ b/packages/core/agent-loop/tests/loop.spec.ts @@ -1,6 +1,6 @@ import { describe, expect, it } from 'vitest' import { Context } from '@deepseek-ai/cordis' -import LlmRuntime, { createUserMessage, CallId, LlmError, StreamChunk } from '@deepseek-ai/dsh-llm' +import LlmRuntime, { createUserMessage, CallId, LlmError, ReasoningEffortId, StreamChunk } from '@deepseek-ai/dsh-llm' import SessionStore, { SessionId, TurnEndReason } from '@deepseek-ai/dsh-session' import SystemPrompt from '@deepseek-ai/dsh-system-prompt' import ToolRuntime, { defineContentToolFixture } from '@deepseek-ai/dsh-tools' @@ -77,6 +77,34 @@ describe('agent loop', () => { expect(adapter.requests[0]?.maxTokens).toBe(256) }) + it('seeds an AgentOptions reasoning effort into the first model request', async () => { + const effort = ReasoningEffortId('high') + const adapter = new MockAdapter([textResponse('reasoned')], { + efforts: [{ id: effort, name: 'High' }], + defaultEffort: effort, + }) + const ctx = await harness(adapter) + const agent = ctx.agentLoop.create( + SessionId('configured-reasoning-effort'), + { provider: 'mock', model: 'mock', reasoningEffort: effort }, + ) + + send(agent, 'use the configured reasoning effort') + await waitForIdle(ctx, agent) + + expect(adapter.requests[0]?.reasoningEffort).toBe(effort) + }) + + it('validates reasoning effort in declarative agent config', () => { + const effort = ReasoningEffortId('high') + expect(AgentLoop.Config({ + agents: [{ id: 'configured-agent', reasoningEffort: effort }], + }).agents[0]?.reasoningEffort).toBe(effort) + expect(() => AgentLoop.Config({ + agents: [{ id: 'configured-agent', reasoningEffort: ReasoningEffortId('') }], + })).toThrow() + }) + it('cancels queued wakeup work together with an active maintenance task', async () => { const adapter = new MockAdapter([textResponse('park reply')]) const ctx = await harness(adapter) @@ -1429,7 +1457,11 @@ describe('agent loop', () => { }) it('creates agents from config on startup', async () => { - const adapter = new MockAdapter([textResponse('from config')]) + const effort = ReasoningEffortId('high') + const adapter = new MockAdapter([textResponse('from config')], { + efforts: [{ id: effort, name: 'High' }], + defaultEffort: effort, + }) const ctx = new Context() await ctx.plugin(LlmRuntime) await ctx.plugin(SessionStore) @@ -1437,7 +1469,7 @@ describe('agent loop', () => { await ctx.plugin(ToolRuntime) await ctx.plugin(AgentRegistry) await ctx.plugin(AgentLoop, { - agents: [{ id: SessionId('config-agent'), provider: 'mock', model: 'mock' }], + agents: [{ id: SessionId('config-agent'), provider: 'mock', model: 'mock', reasoningEffort: effort }], }) ctx.llm.registerAdapter(['mock'], adapter) @@ -1451,6 +1483,9 @@ describe('agent loop', () => { send(agent, 'hi') await waitForIdle(ctx, agent) expect(adapter.requests).toHaveLength(1) + expect(adapter.requests[0]?.reasoningEffort).toBe(effort) + const header = agent.session.events.find(event => event.type === 'request/header') + expect(header?.type === 'request/header' && header.data.header.config.reasoningEffort).toBe(effort) }) it('attaches config agent cwd to the fresh session header', async () => { diff --git a/packages/core/agent/README.i18n.yaml b/packages/core/agent/README.i18n.yaml index 4dd5627b2e..a9d78f3fad 100644 --- a/packages/core/agent/README.i18n.yaml +++ b/packages/core/agent/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/core/agent/README.md -README.md: f6b698e93b254c97786155e7d2c7e81f07c0d981 -README.zh.md: 5e3da8a5d09d06fa4fc9726406aae9c6d5e69c3e +README.md: 70b396d787de5d95332c379ff20ab92c64065857 +README.zh.md: fee72f3cd1fb456ae639d6444fe3fe914c41220a diff --git a/packages/core/agent/README.md b/packages/core/agent/README.md index f6b698e93b..70b396d787 100644 --- a/packages/core/agent/README.md +++ b/packages/core/agent/README.md @@ -14,7 +14,7 @@ Tracks live agents and carries the initiating Agent through asynchronous driver The scoped-registration surface: `Agent.ctx` is the agent's scope context (`dsh-scope`, key = the agent) — register tools/sections/variables/listeners through it for that agent alone, all unwound on disposal. `agentEvents(ctx, agent)` is the fused dispatcher for ordinary agent-subject operations (carrier + injected subject in one move); its notification mode invokes every listener and contains both synchronous throws and returned-promise rejections. The registry lifecycle pair reuses one stable routing carrier. `assembleContextFor(agent)` builds the per-agent assembly context (`agent` + `scope` together). `installModelSelection(agentCtx, selection)` snapshots a mutable provider/model/reasoning-effort selection during prompt assembly, applies its provider and model to prompt variables, and applies the complete selection to request routing for one step; an absent selected effort clears an inherited effort so adapter/provider defaults apply. `CreateAgentOptions.setup(agentCtx)` and `ResumeAgentOptions.setup(agentCtx)` compose a fresh or resumed agent's scoped world while both objects remain unpublished. Setup is trusted, composition-only same-process code: drive the agent only after creation resolves. -`AgentOptions` supplies the initial provider/model route and an optional positive `maxTokens` output cap. The concrete loop resolves any exact-model adapter default, records the effective cap in the request header, and applies it to each conversation-model request; an explicit Agent option wins, while omission leaves the adapter or provider route default in control. +`AgentOptions` supplies the initial provider/model route, optional adapter-owned `reasoningEffort`, and optional positive `maxTokens` output cap. The concrete loop validates exact-model reasoning support, resolves adapter defaults, records the effective values in the request header, and applies them to each conversation-model request; an explicit Agent option wins, while omission leaves the adapter or provider route default in control. - `ctx.agents.register(agent: Agent): () => void` — record an **already-constructed** agent. Disposed with the calling fiber. - Advanced ordered lifecycle: `enter(agent, owner): () => void` enforces `agent.id === agent.session.id`, performs the authoritative ID collision check, and inserts without announcing; `owner` explicitly records the live creator-agent relation (or `undefined` for a root), independently of durable session lineage. `announce(agent)` emits `agent/created` exactly once. A detach requested synchronously by a creation listener is deferred until that dispatch unwinds, and every detach checks the captured entry object, so a stale capability cannot delete a later same-ID replacement. The async factory uses this split; ordinary plugins use `register()`. diff --git a/packages/core/agent/README.zh.md b/packages/core/agent/README.zh.md index 5e3da8a5d0..fee72f3cd1 100644 --- a/packages/core/agent/README.zh.md +++ b/packages/core/agent/README.zh.md @@ -14,7 +14,7 @@ Agent 接口、注册表、进程本地发起方作用域，以及 `agent/*` 事 带作用域的注册接口：`Agent.ctx` 是 agent 的作用域上下文（`dsh-scope`，键 = 该 agent）。通过它注册工具／段／变量／监听器，只对该 agent 生效，并在 dispose（资源释放）时全部撤销。`agentEvents(ctx, agent)` 是普通 agent 主体操作的融合分发器（一次完成载体 + 注入主体）；其通知 mode 会调用每个监听器，并同时收容同步抛出和返回 Promise 的拒绝。注册表生命周期对复用一个稳定路由载体。`assembleContextFor(agent)` 构建按 agent 的组装上下文（同时包含 `agent` + `scope`）。`installAgentLlmTarget(agentCtx, target)` 在提示词组装期间快照可变的提供方／模型／推理（reasoning）强度选择，将路由应用到提示词变量，并将完整目标应用到一个步骤的请求路由；如果没有选定推理强度，则会清除继承的推理强度，使该目标使用适配器／提供方默认值。`CreateAgentOptions.setup(agentCtx)` 和 `ResumeAgentOptions.setup(agentCtx)` 在新建或恢复的 agent 尚未发布时，组合其带作用域的世界。Setup 是受信任、仅用于组合的同进程代码：只有创建完成后才能驱动 agent。 -`AgentOptions` 提供初始的提供方／模型路由，以及可选的正数 `maxTokens` 输出上限。具体循环会解析确切模型的适配器默认值，把生效上限记录到请求 header，并应用到每次对话模型请求；显式 Agent 选项优先，省略时由适配器或提供方路由默认值控制。 +`AgentOptions` 提供初始的提供方／模型路由、可选且由适配器定义的 `reasoningEffort`，以及可选的正数 `maxTokens` 输出上限。具体循环会校验确切模型支持的推理强度、解析适配器默认值，把生效值记录到请求 header，并应用到每次对话模型请求；显式 Agent 选项优先，省略时由适配器或提供方路由默认值控制。 - `ctx.agents.register(agent: Agent): () => void`：记录一个 **已经构造完成** 的 agent。随调用 fiber dispose。 - 高级有序生命周期：`enter(agent, owner): () => void` 强制 `agent.id === agent.session.id`，执行权威 ID 冲突检查，并在不通知的情况下插入；`owner` 显式记录实时创建方 agent 关系（根 agent 为 `undefined`），与持久会话谱系无关。`announce(agent)` 恰好发出一次 `agent/created`。创建监听器同步请求的 detach 会延后到该次分发结束；每次 detach 都会检查捕获的条目对象，因此陈旧能力无法删除后续使用同一 ID 的替代项。异步工厂使用这一拆分；普通插件使用 `register()`。 diff --git a/packages/core/agent/src/runtime-types.ts b/packages/core/agent/src/runtime-types.ts index df7449d406..3f8f7c512b 100644 --- a/packages/core/agent/src/runtime-types.ts +++ b/packages/core/agent/src/runtime-types.ts @@ -7,7 +7,7 @@ import type { Context } from '@deepseek-ai/cordis' import type { Scoped } from '@deepseek-ai/dsh-scope' -import type { LlmCallConfig, LlmFailure, ResolvedRetryPolicy } from '@deepseek-ai/dsh-llm' +import type { LlmCallConfig, LlmFailure, ReasoningEffortId, ResolvedRetryPolicy } from '@deepseek-ai/dsh-llm' import type { AgentCancelCause, Session, UserMessage } from '@deepseek-ai/dsh-session' export type { AgentCancelCause } from '@deepseek-ai/dsh-session' import type { Inbox } from './inbox.ts' @@ -27,6 +27,8 @@ export interface AgentOptions { provider?: string /** Model id interpreted by the selected provider adapter. */ model?: string + /** Adapter-owned reasoning effort for the selected provider/model route. */ + reasoningEffort?: ReasoningEffortId /** Maximum output tokens for each conversation-model request. */ maxTokens?: number } diff --git a/packages/core/session/src/known-event-types.ts b/packages/core/session/src/known-event-types.ts index b4e7117541..c005211567 100644 --- a/packages/core/session/src/known-event-types.ts +++ b/packages/core/session/src/known-event-types.ts @@ -49,6 +49,7 @@ export const KNOWN_SESSION_EVENT_TYPES: ReadonlySet = new Set([ 'step/end', 'step/start', 'subagent/descriptor', + 'subagent/model-selection-enabled', 'team/member', 'team/message/delivered', 'team/message/queued', diff --git a/packages/core/tools/tests/gen-tool-catalog.spec.ts b/packages/core/tools/tests/gen-tool-catalog.spec.ts index ce2007c34b..38d68deae6 100644 --- a/packages/core/tools/tests/gen-tool-catalog.spec.ts +++ b/packages/core/tools/tests/gen-tool-catalog.spec.ts @@ -30,7 +30,7 @@ describe('gen-tool-catalog collectToolCatalog', () => { 'cordis_inspect_query', 'cordis_inspect_self', 'cordis_run', 'cordis_stop', 'cordis_undefine', 'create_goal', 'edit', 'exit_plan_mode', 'followup_task', 'get_goal', 'glob', 'grep', 'interrupt_agent', 'interrupt_agent', 'job_kill', 'job_list', 'job_output', - 'list_agents', 'list_agents', 'lsp', 'pwsh', 'pwsh', 'ralph', + 'list_agents', 'list_agents', 'list_subagent_models', 'lsp', 'pwsh', 'pwsh', 'ralph', 'read', 'read_image', 'report', 'run_code', 'schedule_create', 'schedule_delete', 'schedule_list', 'send_message', 'send_message', 'session_event_read', 'session_event_search', 'session_event_trace', 'session_search', 'session_trace', 'skill', 'spawn_teammate', @@ -88,7 +88,7 @@ describe('gen-tool-catalog collectToolCatalog', () => { // agents surface this one package as both `subagent` and `subagent_fork`. const catalog = await collectToolCatalog() const subagent = catalog.find(entry => entry.pkg === '@deepseek-ai/dsh-tool-subagent') - expect(subagent?.schemas.map(s => s.name)).toEqual(['subagent']) + expect(subagent?.schemas.map(s => s.name)).toEqual(['list_subagent_models', 'subagent']) expect(subagent?.note).toMatch(/subagent_fork/) }) }) diff --git a/packages/experimental/webworker-runtime/tests/fixtures/vfs-example/home/sessions/--dsh-workspace--/preview-architecture-review/session.jsonl b/packages/experimental/webworker-runtime/tests/fixtures/vfs-example/home/sessions/--dsh-workspace--/preview-architecture-review/session.jsonl index 291a18ba1e..7fd433c32a 100644 --- a/packages/experimental/webworker-runtime/tests/fixtures/vfs-example/home/sessions/--dsh-workspace--/preview-architecture-review/session.jsonl +++ b/packages/experimental/webworker-runtime/tests/fixtures/vfs-example/home/sessions/--dsh-workspace--/preview-architecture-review/session.jsonl @@ -171,7 +171,7 @@ {"type":"session/end-seed","data":{},"seq":169,"time":1787472100000} {"type":"turn/start","data":{"turn":29},"seq":170,"time":1787472100001} {"type":"user/message","data":{"id":"preview-review-user","role":"user","content":[{"type":"text","text":"Review whether the preview fixture is isolated from future WebFS data."}],"source":{"kind":"user"}},"surfaceOp":"append","seq":171,"time":1787472100002} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"fork","label":"Review preview architecture"},"seq":172,"time":1787472100003} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"fork","label":"Review preview architecture"},"seq":172,"time":1787472100003} {"type":"step/start","data":{"turn":29,"step":1},"seq":173,"time":1787472100004} {"type":"assistant/message","data":{"turn":29,"step":1,"message":{"id":"preview-review-assistant","role":"assistant","content":[{"type":"text","text":"The bundled fixture is static image content; future WebFS state remains user-owned."}],"source":{"kind":"model","provider":"preview-fixture","model":"deterministic"}}},"sourceEventSeqs":[],"surfaceOp":"append","seq":174,"time":1787472100005} {"type":"step/end","data":{"turn":29,"step":1},"seq":175,"time":1787472100006} diff --git a/packages/experimental/webworker-runtime/tests/fixtures/vfs-example/home/sessions/--dsh-workspace--/preview-follow-up-builder/session.jsonl b/packages/experimental/webworker-runtime/tests/fixtures/vfs-example/home/sessions/--dsh-workspace--/preview-follow-up-builder/session.jsonl index 6353e854c4..03b884adc6 100644 --- a/packages/experimental/webworker-runtime/tests/fixtures/vfs-example/home/sessions/--dsh-workspace--/preview-follow-up-builder/session.jsonl +++ b/packages/experimental/webworker-runtime/tests/fixtures/vfs-example/home/sessions/--dsh-workspace--/preview-follow-up-builder/session.jsonl @@ -1,7 +1,7 @@ {"type":"session","version":0,"id":"preview-follow-up-builder","createdAt":1787472200000,"cwd":"/dsh/workspace","parentSession":"preview-showcase","origin":"subagent","delegationDepth":1,"agentPreset":"standard"} {"type":"turn/start","data":{"turn":1},"seq":0,"time":1787472200000} {"type":"user/message","data":{"id":"preview-builder-user","role":"user","content":[{"type":"text","text":"Check that the Preview workspace can support follow-up tasks."}],"source":{"kind":"user"}},"surfaceOp":"append","seq":1,"time":1787472200001} -{"type":"subagent/descriptor","data":{"version":2,"mode":"continuable","provider":"spawn","label":"Continue preview verification"},"seq":2,"time":1787472200002} +{"type":"subagent/descriptor","data":{"version":3,"mode":"continuable","provider":"spawn","label":"Continue preview verification"},"seq":2,"time":1787472200002} {"type":"step/start","data":{"turn":1,"step":1},"seq":3,"time":1787472200003} {"type":"assistant/message","data":{"turn":1,"step":1,"message":{"id":"preview-builder-assistant","role":"assistant","content":[{"type":"text","text":"This child is continuable and ready for another verification turn."}],"source":{"kind":"model","provider":"preview-fixture","model":"deterministic"}}},"sourceEventSeqs":[],"surfaceOp":"append","seq":4,"time":1787472200004} {"type":"step/end","data":{"turn":1,"step":1},"seq":5,"time":1787472200005} diff --git a/packages/extensions/tool-cordis/src/api-catalog.ts b/packages/extensions/tool-cordis/src/api-catalog.ts index eea3a05e2b..320c44b9aa 100644 --- a/packages/extensions/tool-cordis/src/api-catalog.ts +++ b/packages/extensions/tool-cordis/src/api-catalog.ts @@ -195,7 +195,7 @@ export const SERVICE_API: readonly ServiceApiEntry[] = [ }, { signature: 'async recompose(agentCtx: Context, id: string): Promise', - description: 'Re-link one agent to a different preset\'s standing composition.\n\nOnly valid while the agent has produced nothing: swapping tools mid conversation would leave logged tool calls the new composition cannot make. The CALLER owns that check — this method does not read session history.\n\nThe swap is a parent re-link, not an unmount: standing mounts are shared and permanent, so the old composition stays for its other agents and the new one is ensured BEFORE the link moves. An unknown or unusable preset therefore throws with the agent exactly as it was — there is no torn-down state to restore. The re-link runs through the binding this roster kept from the agent\'s mount — dsh-scope\'s only re-link authority. An agent that never composed one has nothing to re-link: the switch is then the agent\'s first bind, exactly a mount.', + description: 'Re-link one agent to a different preset\'s standing composition.\n\nOnly valid while the agent has produced nothing: swapping tools mid conversation would leave logged tool calls the new composition cannot make. The CALLER owns that check — this method does not read session history.\n\nThe swap is a parent re-link, not an unmount: standing mounts are shared and permanent, so the old composition stays for its other agents and the new one is ensured BEFORE the link moves. An unknown or unusable preset therefore throws with the agent exactly as it was — there is no torn-down state to restore. The re-link runs through the binding this roster kept from the agent\'s mount — dsh-scope\'s only re-link authority. An agent that never composed one has nothing to re-link: the switch is then the agent\'s first bind, exactly a mount. A committed re-link emits `tools/change` because changing the parent scope changes the Agent\'s resolved tool set without adding or removing registry entries.', parameters: [{ name: 'agentCtx', description: 'the agent\'s scope context.' }, { name: 'id', description: 'the preset to compose the agent from instead.' }], returns: 'the preset now installed.', throws: ['when the preset is unknown or its composition is unusable.'], @@ -1906,6 +1906,19 @@ export const SERVICE_API: readonly ServiceApiEntry[] = [ }, ], }, + { + key: 'subagentModelSelection', + summary: 'Singleton settings owner read by delegation tools when an Agent is published.', + description: 'Singleton settings owner read by delegation tools when an Agent is published.', + methods: [ + { + signature: 'currentEnabled(): boolean', + description: 'Read the preference for the next eligible Agent publication.', + parameters: [], + returns: 'whether that Agent should receive model-selectable delegation.', + }, + ], + }, { key: 'subagents', summary: 'Named provider registry with one-shot runs, durable discovery, and continuable-child operations.', @@ -3124,7 +3137,7 @@ export const TYPE_API: readonly TypeApiEntry[] = [ }, { name: 'AgentOptions', - declaration: 'export interface AgentOptions {\n provider?: string;\n model?: string;\n maxTokens?: number;\n}', + declaration: 'export interface AgentOptions {\n provider?: string;\n model?: string;\n reasoningEffort?: ReasoningEffortId;\n maxTokens?: number;\n}', }, { name: 'AgentPreset', @@ -3416,7 +3429,7 @@ export const TYPE_API: readonly TypeApiEntry[] = [ }, { name: 'ContinuableSubagentDescriptorData', - declaration: 'export interface ContinuableSubagentDescriptorData extends SubagentDescriptorBase {\n readonly mode: \'continuable\';\n readonly label: string;\n readonly agentProvider?: string;\n readonly agentModel?: string;\n readonly persona?: string;\n readonly toolFilter?: ToolRestriction;\n}', + declaration: 'export interface ContinuableSubagentDescriptorData extends SubagentDescriptorBase {\n readonly mode: \'continuable\';\n readonly label: string;\n readonly agentProvider?: string;\n readonly agentModel?: string;\n readonly agentReasoningEffort?: ReasoningEffortId;\n readonly persona?: string;\n readonly toolFilter?: ToolRestriction;\n}', }, { name: 'CordisDynamicPackageId', @@ -4928,7 +4941,7 @@ export const TYPE_API: readonly TypeApiEntry[] = [ }, { name: 'SubagentCapabilities', - declaration: 'export interface SubagentCapabilities {\n readonly outputSchema: boolean;\n readonly depthLimit: boolean;\n readonly toolFilter: boolean;\n readonly persona: boolean;\n}', + declaration: 'export interface SubagentCapabilities {\n readonly agentOptions: boolean;\n readonly outputSchema: boolean;\n readonly depthLimit: boolean;\n readonly toolFilter: boolean;\n readonly persona: boolean;\n}', }, { name: 'SubagentDescendantListEntry', diff --git a/packages/llm/llm-deepseek/README.i18n.yaml b/packages/llm/llm-deepseek/README.i18n.yaml index 2d2b710299..57ecda0a84 100644 --- a/packages/llm/llm-deepseek/README.i18n.yaml +++ b/packages/llm/llm-deepseek/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/llm/llm-deepseek/README.md -README.md: 0570ead15e0c408e838cab7a641293ccdb11a702 -README.zh.md: 7f58ba2b6a9be8f03fdcc4538777889780949057 +README.md: 7433bb75104506ec2409c659f3d30058abc6f9a4 +README.zh.md: 7dcdfeac17b0bfca70a293760061182292edb531 diff --git a/packages/llm/llm-deepseek/README.md b/packages/llm/llm-deepseek/README.md index 0570ead15e..7433bb7510 100644 --- a/packages/llm/llm-deepseek/README.md +++ b/packages/llm/llm-deepseek/README.md @@ -50,7 +50,7 @@ The package root exposes the Cordis plugin contract and `DeepSeekAdapter`; wire contextWindow: 512000 ``` -The plugin registers the single provider route `deepseek-official` together with its resolved `retryPolicy`; omission resolves to normal mode with five retries. A request selects it with `provider: deepseek-official`; its `model` is passed through as the wire `model` string, so changing DeepSeek models does not require lifecycle-time registration. Omitting `models` advertises `deepseek-v4-flash`, `deepseek-v4-pro`, and the image-capable `deepseek-v4-flash-vision-exp`, each with a 1,000,000-token context window; an explicit list replaces those defaults, while `models: []` advertises none. Catalog entries are exposed through `ctx.llm.listModels('deepseek-official')` for clients such as ACP editors and the Web selector, but remain advisory: unlisted model ids still pass through unchanged as text-only routes. An omitted entry name defaults to its id, and omitted `inputModalities` means `text` only. +The plugin registers the single provider route `deepseek-official` together with its resolved `retryPolicy`; omission resolves to normal mode with five retries. A request selects it with `provider: deepseek-official`; its `model` is passed through as the wire `model` string, so changing DeepSeek models does not require lifecycle-time registration. Omitting `models` advertises `deepseek-v4-flash` as the fast, economical choice for focused work, `deepseek-v4-pro` as the stronger, higher-cost choice for complex or quality-critical work, and the image-capable `deepseek-v4-flash-vision-exp`; each has a 1,000,000-token context window. An explicit list replaces those defaults, while `models: []` advertises none. Catalog entries are exposed through `ctx.llm.listModels('deepseek-official')` for clients such as ACP editors, the Web selector, and model discovery tools, but remain advisory: unlisted model ids still pass through unchanged as text-only routes. An omitted entry name defaults to its id, and omitted `inputModalities` means `text` only. An image-capable catalog entry declares `inputModalities: [text, image]` and may set `imagePixelBudget` to an exact positive integer or `low`; omission uses 640,000 total pixels, while `low` selects 512 by 512 total pixels. `imageMaxBytes` defaults to 1MiB. The attachment store scales by `min(1, sqrt(pixelBudget / (width * height)))` and rounds inward to keep the pixel count at or below the hard cap, so a 2048 by 1024 normalized attachment becomes about 1130 by 565 instead of a forced square. Request encoders run lazily: alpha images try WebP (effort 0) at 85, 75, then 60, and opaque images try JPEG at those qualities; when every quality exceeds 1MiB the smallest output is used. Concurrent generation of one `variantId` shares one transform. A caller can cancel its own wait without interrupting other waiters; the transform stops when no waiter remains. The adapter normally uploads the exact derived request bytes through `POST /files` and sends `{type: "file", file_id}` blocks. A failed or timed-out file-id resolution rebuilds the whole chat request with those same request versions as base64 data URLs; one request never mixes file ids and inline images. Every retained image is preceded by text naming the complete attachment id and actual request dimensions. When the attachment provider exposes a host object and the current filesystem maps it into the tool execution world, the text also includes that read-only path and the matching extension for a writable copy. This access is resolved independently from the deterministic request version and its `variantId`. The descriptor states that the preview and normalized image may differ from the upload. User, tool-result, agent-loop, compaction, and direct `ctx.llm.stream` requests all use this projection. Text-only routes receive stable attachment placeholders while durable history keeps its image references. @@ -66,7 +66,7 @@ Concurrent resolution of one scoped `variantId` shares one Files upload with wai `maxTokens` is the adapter-configured output cap for conversation requests and defaults to 256,000. A catalog entry may carry its own `maxTokens`, which wins for that model; an entry without one, and any unlisted pass-through id, resolve to the profile value, so adding a per-model cap changes one model rather than the route. Exact-model resolution exposes the winner as `defaultMaxTokens`; `LlmRuntime` materializes that value into `GenerateOptions.maxTokens` before the agent loop writes `request/header`, so the wire request remains reconstructable. An explicit request or `AgentOptions.maxTokens` value wins and is serialized as `max_tokens`. The adapter does not clamp this request budget against `contextWindow`; deployments with a smaller context or provider output limit must configure a compatible `maxTokens`. -The same exact-model result exposes ordered `off`, `low`, `high`, and `max` efforts under `reasoning` for every pass-through model when deployment policy permits thinking. `reasoningEffort` selects the deployment default and falls back to `high` when omitted. `agent/request` can replace it on each conversation step; the resolved value is logged in `request/header`. `low`, `high`, and `max` enable thinking and serialize as the same official top-level `reasoning_effort` value; adapter-owned `off` instead serializes `thinking.type: disabled` and omits `reasoning_effort`. An unsupported value fails with `UNSUPPORTED_REASONING_EFFORT` before network I/O. +The same exact-model result exposes ordered `off`, `low`, `high`, and `max` efforts with selection guidance under `reasoning` for every pass-through model when deployment policy permits thinking. `reasoningEffort` selects the advertised default and falls back to `high` when omitted. `agent/request` can replace it on each conversation step; the resolved value is logged in `request/header`. `low`, `high`, and `max` enable thinking and serialize as the same official top-level `reasoning_effort` value; adapter-owned `off` instead serializes `thinking.type: disabled` and omits `reasoning_effort`. An unsupported value fails with `UNSUPPORTED_REASONING_EFFORT` before network I/O. `thinking: disabled` is a deployment lock that publishes only `off` with `off` as its default. Omitting `reasoningEffort` or configuring it as `off` is valid; configuring `low`, `high`, or `max` fails plugin loading, and a direct per-request attempt to enable thinking fails before network I/O. A request with `GenerateOptions.purpose: 'session-title'` also forces thinking disabled and omits the already-resolved effort, reserving its bounded output for visible title text without changing conversation or compaction defaults. diff --git a/packages/llm/llm-deepseek/README.zh.md b/packages/llm/llm-deepseek/README.zh.md index 7f58ba2b6a..7dcdfeac17 100644 --- a/packages/llm/llm-deepseek/README.zh.md +++ b/packages/llm/llm-deepseek/README.zh.md @@ -50,7 +50,7 @@ harness LLM（大语言模型）seam 的 DeepSeek chat-completions 适配器： contextWindow: 512000 ``` -该插件注册唯一提供方路由 `deepseek-official`，并一同注册解析后的 `retryPolicy`；省略时会解析为 normal 模式并重试五次。请求使用 `provider: deepseek-official` 选择该路由；其 `model` 会作为协议 `model` 字符串原样传递，因此更改 DeepSeek 模型不需要生命周期时注册。省略 `models` 会公布 `deepseek-v4-flash`、`deepseek-v4-pro` 与支持图片输入的 `deepseek-v4-flash-vision-exp`，三者的上下文窗口均为 1,000,000 token；显式列表会替换这些默认值，`models: []` 则不公布任何模型。Catalog 配置项通过 `ctx.llm.listModels('deepseek-official')` 公开给 ACP（Agent Client Protocol）编辑器和 Web 选择器等客户端，但仍只提供建议：未列出模型 id 仍原样传递，并按纯文本路由处理。省略配置项 name 默认为其 id，省略 `inputModalities` 则表示仅支持 `text`。 +该插件注册唯一提供方路由 `deepseek-official`，并一同注册解析后的 `retryPolicy`；省略时会解析为 normal 模式并重试五次。请求使用 `provider: deepseek-official` 选择该路由；其 `model` 会作为协议 `model` 字符串原样传递，因此更改 DeepSeek 模型不需要生命周期时注册。省略 `models` 会公布适合聚焦任务、快速且经济的 `deepseek-v4-flash`，适合复杂或质量关键任务、更强但成本更高的 `deepseek-v4-pro`，以及支持图片输入的 `deepseek-v4-flash-vision-exp`；三者的上下文窗口均为 1,000,000 token。显式列表会替换这些默认值，`models: []` 则不公布任何模型。Catalog 配置项通过 `ctx.llm.listModels('deepseek-official')` 公开给 ACP（Agent Client Protocol）编辑器、Web 选择器和模型发现工具等客户端，但仍只提供建议：未列出模型 id 仍原样传递，并按纯文本路由处理。省略配置项 name 默认为其 id，省略 `inputModalities` 则表示仅支持 `text`。 支持图片的 catalog 配置项声明 `inputModalities: [text, image]`，并可把 `imagePixelBudget` 设为确切正整数或 `low`；省略时使用总像素 640,000，`low` 选择总像素 512×512。`imageMaxBytes` 默认值为 1MiB。附件存储按 `min(1, sqrt(pixelBudget / (width * height)))` 缩放，并向预算内取整，确保总像素不超过硬上限。因此 2048×1024 规范化附件会得到约 1130×565 的请求版本，而不会被强制变成正方形。请求编码按需执行：透明图片依次尝试质量 85、75、60 的 WebP（effort 0）；非透明图片依次尝试这些质量的 JPEG。全部质量档都超过 1MiB 时使用其中最小的产物。同一 `variantId` 的并发生成共享一次变换。调用方可以单独取消等待，不会中断其他等待方；没有等待方时才会停止变换。适配器通常通过 `POST /files` 上传确切的派生请求字节，再发送 `{type: "file", file_id}` 块。File ID 解析失败或超时后，适配器会用相同请求版本的 base64 data URL 重新组装整个 chat 请求；同一请求不会混用 file ID 和内联图片。每张保留图片前都有文本，写明完整附件 ID 和实际请求尺寸。附件提供方给出宿主对象且当前文件系统能够将其映射到工具执行环境时，文本还会给出该只读路径，并指出复制到可写路径时应使用的匹配扩展名。该访问方式独立于确定性的请求版本及其 `variantId`。描述也会说明预览和规范化图片可能与上传图片不同。User、工具结果、agent loop、压缩和直接 `ctx.llm.stream` 请求都使用该投影。纯文本路由会收到稳定的附件占位文本，持久历史继续保留图片引用。 @@ -66,7 +66,7 @@ harness LLM（大语言模型）seam 的 DeepSeek chat-completions 适配器： `maxTokens` 是适配器为对话请求配置的输出上限，默认值为 256,000。Catalog 配置项可以自带 `maxTokens`，它对该模型胜出；不含该上限的配置项以及任何未列出原样传递 id 都解析为 profile 值，因此新增按模型的上限只改变一个模型，而非整条路由。确切模型解析会将胜出值公开为 `defaultMaxTokens`；`LlmRuntime` 会在 agent loop（智能体循环）写入 `request/header` 前，将该值填入 `GenerateOptions.maxTokens`，从而仍可根据持久记录重建协议请求。显式的请求值或 `AgentOptions.maxTokens` 值优先，并会序列化为 `max_tokens`。适配器不会根据 `contextWindow` 自动调低该请求预算；上下文或提供方输出上限较小的部署必须配置与其相容的 `maxTokens`。 -同一确切模型结果会在部署策略允许思考时，为每个原样传递模型在 `reasoning` 下公开有序的 `off`、`low`、`high` 和 `max` 推理（reasoning）强度。`reasoningEffort` 选择部署默认值，省略时回退为 `high`。`agent/request` 可以在每个会话步骤替换它；解析后的值会记录在 `request/header`。`low`、`high` 和 `max` 会启用思考，并以同名值序列化为官方顶层 `reasoning_effort`；适配器持有的 `off` 则序列化为 `thinking.type: disabled`，且省略 `reasoning_effort`。不支持的值会在网络 I/O 前以 `UNSUPPORTED_REASONING_EFFORT` 失败。 +同一确切模型结果会在部署策略允许思考时，为每个原样传递模型在 `reasoning` 下公开有序且附带选择指引的 `off`、`low`、`high` 和 `max` 推理（reasoning）强度。`reasoningEffort` 选择公布的默认值，省略时回退为 `high`。`agent/request` 可以在每个会话步骤替换它；解析后的值会记录在 `request/header`。`low`、`high` 和 `max` 会启用思考，并以同名值序列化为官方顶层 `reasoning_effort`；适配器持有的 `off` 则序列化为 `thinking.type: disabled`，且省略 `reasoning_effort`。不支持的值会在网络 I/O 前以 `UNSUPPORTED_REASONING_EFFORT` 失败。 `thinking: disabled` 是部署锁定：它只公布 `off`，并以 `off` 为默认值。省略 `reasoningEffort` 或将其配置为 `off` 均有效；配置 `low`、`high` 或 `max` 会使插件加载失败，直接按请求启用思考也会在网络 I/O 前失败。携带 `GenerateOptions.purpose: 'session-title'` 的请求也会强制禁用思考并省略已解析的推理强度，将有界输出保留给可见标题文本，不改变会话或压缩（compaction）默认值。 diff --git a/packages/llm/llm-deepseek/src/adapter.ts b/packages/llm/llm-deepseek/src/adapter.ts index 92761bb840..35212d32d6 100644 --- a/packages/llm/llm-deepseek/src/adapter.ts +++ b/packages/llm/llm-deepseek/src/adapter.ts @@ -173,13 +173,33 @@ const LOW_REASONING_EFFORT = ReasoningEffortId('low') const HIGH_REASONING_EFFORT = ReasoningEffortId('high') const MAX_REASONING_EFFORT = ReasoningEffortId('max') const REASONING_EFFORTS = [ - { id: OFF_REASONING_EFFORT, name: 'Off' }, - { id: LOW_REASONING_EFFORT, name: 'Low' }, - { id: HIGH_REASONING_EFFORT, name: 'High' }, - { id: MAX_REASONING_EFFORT, name: 'Max' }, + { + id: OFF_REASONING_EFFORT, + name: 'Off', + description: 'Use for simple tasks that do not need reasoning.', + }, + { + id: LOW_REASONING_EFFORT, + name: 'Low', + description: 'Prefer for routine or latency-sensitive tasks.', + }, + { + id: HIGH_REASONING_EFFORT, + name: 'High', + description: 'The default balance for most tasks.', + }, + { + id: MAX_REASONING_EFFORT, + name: 'Max', + description: 'Reserve for the hardest quality-first tasks.', + }, ] as const const OFF_ONLY_REASONING_EFFORTS = [ - { id: OFF_REASONING_EFFORT, name: 'Off' }, + { + id: OFF_REASONING_EFFORT, + name: 'Off', + description: 'Use for simple tasks that do not need reasoning.', + }, ] as const /** Marks a failed file-id resolution that may be retried as an inline request. */ diff --git a/packages/llm/llm-deepseek/src/index.ts b/packages/llm/llm-deepseek/src/index.ts index 0d895a5e3b..da0c9e24b5 100644 --- a/packages/llm/llm-deepseek/src/index.ts +++ b/packages/llm/llm-deepseek/src/index.ts @@ -82,8 +82,18 @@ const DEFAULT_API_KEY_ENV = 'DEEPSEEK_API_KEY' const PROVIDER = 'deepseek-official' const DEFAULT_MODELS: DeepSeekCatalogModel[] = [ - { id: 'deepseek-v4-flash', name: 'DeepSeek-V4-Flash', contextWindow: DEFAULT_CONTEXT_WINDOW }, - { id: 'deepseek-v4-pro', name: 'DeepSeek-V4-Pro', contextWindow: DEFAULT_CONTEXT_WINDOW }, + { + id: 'deepseek-v4-flash', + name: 'DeepSeek-V4-Flash', + description: 'Fast, efficient, and economical; suited to focused, routine, or parallel tasks.', + contextWindow: DEFAULT_CONTEXT_WINDOW, + }, + { + id: 'deepseek-v4-pro', + name: 'DeepSeek-V4-Pro', + description: 'Stronger agentic coding, knowledge, and difficult reasoning; suited to complex or quality-critical tasks at higher cost.', + contextWindow: DEFAULT_CONTEXT_WINDOW, + }, { id: 'deepseek-v4-flash-vision-exp', name: 'DeepSeek-V4-Flash-Vision-Exp', diff --git a/packages/llm/llm-deepseek/tests/adapter.spec.ts b/packages/llm/llm-deepseek/tests/adapter.spec.ts index 9366c32ce8..3f87737930 100644 --- a/packages/llm/llm-deepseek/tests/adapter.spec.ts +++ b/packages/llm/llm-deepseek/tests/adapter.spec.ts @@ -1205,7 +1205,11 @@ describe('DeepSeekAdapter against a mock server', () => { await expect(ctx.llm.resolveModelInfo('deepseek-official', 'deepseek-v4-flash')) .resolves.toMatchObject({ reasoning: { - efforts: [{ id: ReasoningEffortId('off'), name: 'Off' }], + efforts: [{ + id: ReasoningEffortId('off'), + name: 'Off', + description: 'Use for simple tasks that do not need reasoning.', + }], defaultEffort: ReasoningEffortId('off'), }, }) @@ -1665,8 +1669,20 @@ describe('plugin registration and config', () => { await ctx.plugin(LlmDeepSeek, { baseURL: 'http://127.0.0.1:1' }) expect(ctx.llm.listProviders()).toEqual([{ id: 'deepseek-official', name: 'DeepSeek' }]) await expect(ctx.llm.listModels('deepseek-official')).resolves.toEqual([ - { provider: 'deepseek-official', id: 'deepseek-v4-flash', name: 'DeepSeek-V4-Flash', inputModalities: ['text'] }, - { provider: 'deepseek-official', id: 'deepseek-v4-pro', name: 'DeepSeek-V4-Pro', inputModalities: ['text'] }, + { + provider: 'deepseek-official', + id: 'deepseek-v4-flash', + name: 'DeepSeek-V4-Flash', + description: 'Fast, efficient, and economical; suited to focused, routine, or parallel tasks.', + inputModalities: ['text'], + }, + { + provider: 'deepseek-official', + id: 'deepseek-v4-pro', + name: 'DeepSeek-V4-Pro', + description: 'Stronger agentic coding, knowledge, and difficult reasoning; suited to complex or quality-critical tasks at higher cost.', + inputModalities: ['text'], + }, { provider: 'deepseek-official', id: 'deepseek-v4-flash-vision-exp', name: 'DeepSeek-V4-Flash-Vision-Exp', inputModalities: ['text', 'image'] }, ]) await expect(ctx.llm.resolveModelInfo('deepseek-official', 'deepseek-v4-flash')) @@ -1678,10 +1694,10 @@ describe('plugin registration and config', () => { defaultMaxTokens: 256_000, reasoning: { efforts: [ - { id: ReasoningEffortId('off'), name: 'Off' }, - { id: ReasoningEffortId('low'), name: 'Low' }, - { id: ReasoningEffortId('high'), name: 'High' }, - { id: ReasoningEffortId('max'), name: 'Max' }, + { id: ReasoningEffortId('off'), name: 'Off', description: 'Use for simple tasks that do not need reasoning.' }, + { id: ReasoningEffortId('low'), name: 'Low', description: 'Prefer for routine or latency-sensitive tasks.' }, + { id: ReasoningEffortId('high'), name: 'High', description: 'The default balance for most tasks.' }, + { id: ReasoningEffortId('max'), name: 'Max', description: 'Reserve for the hardest quality-first tasks.' }, ], defaultEffort: ReasoningEffortId('high'), }, @@ -1708,10 +1724,10 @@ describe('plugin registration and config', () => { .resolves.toMatchObject({ reasoning: { efforts: [ - { id: ReasoningEffortId('off'), name: 'Off' }, - { id: ReasoningEffortId('low'), name: 'Low' }, - { id: ReasoningEffortId('high'), name: 'High' }, - { id: ReasoningEffortId('max'), name: 'Max' }, + { id: ReasoningEffortId('off'), name: 'Off', description: 'Use for simple tasks that do not need reasoning.' }, + { id: ReasoningEffortId('low'), name: 'Low', description: 'Prefer for routine or latency-sensitive tasks.' }, + { id: ReasoningEffortId('high'), name: 'High', description: 'The default balance for most tasks.' }, + { id: ReasoningEffortId('max'), name: 'Max', description: 'Reserve for the hardest quality-first tasks.' }, ], defaultEffort: ReasoningEffortId(effort), }, @@ -1729,7 +1745,11 @@ describe('plugin registration and config', () => { await expect(ctx.llm.resolveModelInfo('deepseek-official', 'unlisted-pass-through')) .resolves.toMatchObject({ reasoning: { - efforts: [{ id: ReasoningEffortId('off'), name: 'Off' }], + efforts: [{ + id: ReasoningEffortId('off'), + name: 'Off', + description: 'Use for simple tasks that do not need reasoning.', + }], defaultEffort: ReasoningEffortId('off'), }, }) @@ -1761,7 +1781,11 @@ describe('plugin registration and config', () => { const adapter = adapterOf({ thinking: 'disabled', reasoningEffort: 'off' }) await expect(adapter.resolveModel('deepseek-official', 'pass-through')).resolves.toMatchObject({ reasoning: { - efforts: [{ id: ReasoningEffortId('off'), name: 'Off' }], + efforts: [{ + id: ReasoningEffortId('off'), + name: 'Off', + description: 'Use for simple tasks that do not need reasoning.', + }], defaultEffort: ReasoningEffortId('off'), }, }) @@ -1772,8 +1796,20 @@ describe('plugin registration and config', () => { await ctx.plugin(LlmRuntime) LlmDeepSeek.apply(ctx, { baseURL: 'http://127.0.0.1:1' }) await expect(ctx.llm.listModels('deepseek-official')).resolves.toEqual([ - { provider: 'deepseek-official', id: 'deepseek-v4-flash', name: 'DeepSeek-V4-Flash', inputModalities: ['text'] }, - { provider: 'deepseek-official', id: 'deepseek-v4-pro', name: 'DeepSeek-V4-Pro', inputModalities: ['text'] }, + { + provider: 'deepseek-official', + id: 'deepseek-v4-flash', + name: 'DeepSeek-V4-Flash', + description: 'Fast, efficient, and economical; suited to focused, routine, or parallel tasks.', + inputModalities: ['text'], + }, + { + provider: 'deepseek-official', + id: 'deepseek-v4-pro', + name: 'DeepSeek-V4-Pro', + description: 'Stronger agentic coding, knowledge, and difficult reasoning; suited to complex or quality-critical tasks at higher cost.', + inputModalities: ['text'], + }, { provider: 'deepseek-official', id: 'deepseek-v4-flash-vision-exp', name: 'DeepSeek-V4-Flash-Vision-Exp', inputModalities: ['text', 'image'] }, ]) }) diff --git a/packages/preset/agent-presets/README.i18n.yaml b/packages/preset/agent-presets/README.i18n.yaml index 42601edbd4..a4f3fed8d6 100644 --- a/packages/preset/agent-presets/README.i18n.yaml +++ b/packages/preset/agent-presets/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/preset/agent-presets/README.md -README.md: e204868e43ba07b78eff31293cc33cf790ff0085 -README.zh.md: f11fa5f092b9d4de1abb5c1464b6931d749c2970 +README.md: 3e92f02518c5a36412c9448a26d32958c217f79c +README.zh.md: 4600589749a063df924f9c961cc449506ba2af9f diff --git a/packages/preset/agent-presets/README.md b/packages/preset/agent-presets/README.md index e204868e43..3e92f02518 100644 --- a/packages/preset/agent-presets/README.md +++ b/packages/preset/agent-presets/README.md @@ -16,7 +16,7 @@ Discovery is unmemoized: `list()` and `resolve()` re-read the roots on every cal - `ctx.agentPresets.mount(agentCtx, id?): Promise` Compose one agent from a preset — ensure its standing mount (single-flight) and parent the agent's scope key to it — returning the preset for the caller to record. Refuses a broken preset up front with its discovery-reported reason, so every unloadable shape fails the same way before the loader is involved. - `ctx.agentPresets.composeFrom(agentCtx, parentCtx): string | undefined` Join one agent to the standing composition another already runs on, returning the preset id joined — `undefined` when the parent joined none, which is the rosterless deployment and not an error. A bind rather than a mount, so it is synchronous and has no composition failure mode; it still rejects a caller error (an unscoped context, or an agent that already joined). - `ctx.agentPresets.composedPreset(agentCtx): string | undefined` The preset one LIVE agent runs on, read from its scope chain rather than from its session — the only answer available for an agent whose durable header is still being built. -- `ctx.agentPresets.recompose(agentCtx, id): Promise` Re-link one agent to a different preset's standing composition. Valid only while the agent has produced nothing — **the caller owns that check**; the new mount is ensured before the link moves, so a failure leaves the agent as it was. Refuses a broken preset like `mount()`. +- `ctx.agentPresets.recompose(agentCtx, id): Promise` Re-link one agent to a different preset's standing composition. Valid only while the agent has produced nothing — **the caller owns that check**; the new mount is ensured before the link moves, so a failure leaves the agent as it was. A committed re-link emits `tools/change`, because the Agent's resolved tool set changed without a registry entry changing. Refuses a broken preset like `mount()`. - `ctx.agentPresets.standingKeyFor(id?): Promise` The standing scope key a host reader with no agent (a cold transcript read) resolves preset registrations in; ensures the mount without starting an agent, session, or turn. Refuses a broken preset like `mount()`. - `ctx.agentPresets.roots: readonly PresetRoot[]` The roots this roster scans — every configured root in order, then the derived harness-home root. Not `config.roots`: read this to answer whether a roster is composed at all, so one derivation decides it. - `ctx.agentPresets.authorable: boolean` Whether any of those roots has `user` trust, and therefore whether a preset can be created at all. diff --git a/packages/preset/agent-presets/README.zh.md b/packages/preset/agent-presets/README.zh.md index f11fa5f092..4600589749 100644 --- a/packages/preset/agent-presets/README.zh.md +++ b/packages/preset/agent-presets/README.zh.md @@ -16,7 +16,7 @@ - `ctx.agentPresets.mount(agentCtx, id?): Promise` 用一个 preset 组装一个 agent——确保其常驻挂载（并发去重）并把 agent 的 scope key 认父到它——返回该 preset 供调用方记录。对损坏的 preset 直接以发现时记下的原因拒绝，所以每种不可加载的形态都在加载器介入之前以同一方式失败。 - `ctx.agentPresets.composeFrom(agentCtx, parentCtx): string | undefined` 让一个 agent 加入另一个 agent 已在运行的常驻组装，返回所加入的 preset id——父方未加入任何 preset 时返回 `undefined`，那是无 roster 的部署，不是错误。这是认父而非挂载，因此同步、且自身没有组装失败模式；调用方用错（上下文无 scope、agent 已加入过）仍会拒绝。 - `ctx.agentPresets.composedPreset(agentCtx): string | undefined` 某个**活着的** agent 正在运行的 preset，从其 scope 链读取而不是从其会话读取——对于持久化 header 尚在构建中的 agent，这是唯一能拿到的答案。 -- `ctx.agentPresets.recompose(agentCtx, id): Promise` 把一个 agent 重链到另一个 preset 的常驻组装。仅在该 agent 尚无任何产出时合法——**由调用方负责该检查**；新挂载在链移动之前确保完成，失败时 agent 原封不动。与 `mount()` 一样拒绝损坏的 preset。 +- `ctx.agentPresets.recompose(agentCtx, id): Promise` 把一个 agent 重链到另一个 preset 的常驻组装。仅在该 agent 尚无任何产出时合法——**由调用方负责该检查**；新挂载在链移动之前确保完成，失败时 agent 原封不动。重链提交后会发出 `tools/change`，因为 Agent 解析到的工具集已经变化、但注册表条目本身没有增删。与 `mount()` 一样拒绝损坏的 preset。 - `ctx.agentPresets.standingKeyFor(id?): Promise` 没有 agent 的宿主读取方（冷读记录）解析 preset 注册所用的常驻 scope key；确保挂载而不启动任何 agent、会话或轮次。与 `mount()` 一样拒绝损坏的 preset。 - `ctx.agentPresets.roots: readonly PresetRoot[]` 本 roster 实际扫描的根目录——全部已配置根目录按序在前，随后是推导出的 harness home 根目录。它不是 `config.roots`：判断「是否已组装 roster」应读它，从而由同一处推导决定。 - `ctx.agentPresets.authorable: boolean` 上述根目录中是否有任一具备 `user` 信任级别，因而 preset 是否可创建。 diff --git a/packages/preset/agent-presets/package.json b/packages/preset/agent-presets/package.json index dfdf08cbbe..a20ab0a87c 100644 --- a/packages/preset/agent-presets/package.json +++ b/packages/preset/agent-presets/package.json @@ -48,6 +48,7 @@ "@deepseek-ai/dsh-session": "workspace:^", "@deepseek-ai/dsh-settings": "workspace:^", "@deepseek-ai/dsh-system-prompt": "workspace:^", + "@deepseek-ai/dsh-tools": "workspace:^", "@deepseek-ai/cordis": "workspace:^" }, "dependencies": { diff --git a/packages/preset/agent-presets/presets/code/agent.cordis.yml b/packages/preset/agent-presets/presets/code/agent.cordis.yml index 3333a980c0..1302329c25 100644 --- a/packages/preset/agent-presets/presets/code/agent.cordis.yml +++ b/packages/preset/agent-presets/presets/code/agent.cordis.yml @@ -189,8 +189,13 @@ config: provider: spawn toolName: subagent + modelSelectionSettings: true backgroundMode: continuable + # Fork omits model selection so provider/model stay equal to the parent and + # the inherited history remains eligible for KV Cache reuse. This preset + # keeps fork continuable and accepts its child-scoped `report` additions invalidating + # that prefix; issue #2124 tracks cache-preserving continuable fork. - id: tool-subagent-fork name: '@deepseek-ai/dsh-tool-subagent' config: diff --git a/packages/preset/agent-presets/presets/cordis/agent.cordis.yml b/packages/preset/agent-presets/presets/cordis/agent.cordis.yml index f23907c655..b016eae3a1 100644 --- a/packages/preset/agent-presets/presets/cordis/agent.cordis.yml +++ b/packages/preset/agent-presets/presets/cordis/agent.cordis.yml @@ -176,8 +176,13 @@ config: provider: spawn toolName: subagent + modelSelectionSettings: true backgroundMode: continuable + # Fork omits model selection so provider/model stay equal to the parent and + # the inherited history remains eligible for KV Cache reuse. This preset + # keeps fork continuable and accepts its child-scoped `report` additions invalidating + # that prefix; issue #2124 tracks cache-preserving continuable fork. - id: tool-subagent-fork name: '@deepseek-ai/dsh-tool-subagent' config: diff --git a/packages/preset/agent-presets/presets/standard/agent.cordis.yml b/packages/preset/agent-presets/presets/standard/agent.cordis.yml index 5cb19e1e24..c21c5e4d79 100644 --- a/packages/preset/agent-presets/presets/standard/agent.cordis.yml +++ b/packages/preset/agent-presets/presets/standard/agent.cordis.yml @@ -188,8 +188,13 @@ config: provider: spawn toolName: subagent + modelSelectionSettings: true backgroundMode: continuable + # Fork omits model selection so provider/model stay equal to the parent and + # the inherited history remains eligible for KV Cache reuse. This preset + # keeps fork continuable and accepts its child-scoped `report` additions invalidating + # that prefix; issue #2124 tracks cache-preserving continuable fork. - id: tool-subagent-fork name: '@deepseek-ai/dsh-tool-subagent' config: diff --git a/packages/preset/agent-presets/src/index.ts b/packages/preset/agent-presets/src/index.ts index e22dee7275..8f44e2003c 100644 --- a/packages/preset/agent-presets/src/index.ts +++ b/packages/preset/agent-presets/src/index.ts @@ -27,6 +27,8 @@ import z from '@deepseek-ai/schemastery' import { bindScopeParent, createScope, scopeOf, type Scope, type ScopeKey, type ScopeParentBinding } from '@deepseek-ai/dsh-scope' // Type-only: resolves the `agent/created` lifecycle event this service watches. import type {} from '@deepseek-ai/dsh-agent' +// Type-only: resolves the registry notification emitted after scope reparenting. +import type {} from '@deepseek-ai/dsh-tools' import { settingsNamespace, type SettingsScope, type default as SettingsService } from '@deepseek-ai/dsh-settings' import { dshHomePath } from '@deepseek-ai/dsh-home-paths' import { discoverPresets, SHIPPED_PRESET_ROOT, USER_PRESET_DIR } from './discovery.ts' @@ -455,7 +457,9 @@ export class AgentPresets extends Service { * state to restore. The re-link runs through the binding this roster kept * from the agent's mount — dsh-scope's only re-link authority. An agent * that never composed one has nothing to re-link: the switch is then the - * agent's first bind, exactly a mount. + * agent's first bind, exactly a mount. A committed re-link emits + * `tools/change` because changing the parent scope changes the Agent's + * resolved tool set without adding or removing registry entries. * @param agentCtx - the agent's scope context. * @param id - the preset to compose the agent from instead. * @returns the preset now installed. @@ -474,6 +478,14 @@ export class AgentPresets extends Service { } else { binding.rebind(standing.key) } + // Reparenting changes every scope-layered tool view without adding or + // removing a registration. Publish the registry's normal invalidation so + // Agent-owned overlays can reconcile with the new ancestry. + try { + this.ctx.emit('tools/change') + } catch (error: unknown) { + this.ctx.logger.warn(`agent-presets: tools/change listener failed after recomposing an Agent: ${String(error)}`) + } return preset } diff --git a/packages/preset/agent-presets/tests/mount.spec.ts b/packages/preset/agent-presets/tests/mount.spec.ts index 5b22e26ec6..83f6da3102 100644 --- a/packages/preset/agent-presets/tests/mount.spec.ts +++ b/packages/preset/agent-presets/tests/mount.spec.ts @@ -484,6 +484,27 @@ describe('replacing a composition', () => { expect(toolNames(ctx)).toEqual([]) }) + it('notifies tool views after reparenting and contains notification failures', async () => { + const handle = await ctx.agents.create({ + sessionId: SessionId('sess-tool-change'), + setup: async (agentCtx: Context) => void await ctx.agentPresets.mount(agentCtx, 'standard'), + }) + await ctx.agentPresets.standingKeyFor('minimal') + let changes = 0 + const stopCounting = ctx.on('tools/change', () => { changes += 1 }) + await ctx.agentPresets.recompose(handle.agent.ctx, 'minimal') + expect(changes).toBe(1) + stopCounting() + + const warnings: string[] = [] + ctx.logger.warn = ((message: unknown) => { warnings.push(String(message)) }) as typeof ctx.logger.warn + const stopThrowing = ctx.on('tools/change', () => { throw new Error('listener failed') }) + await expect(ctx.agentPresets.recompose(handle.agent.ctx, 'standard')).resolves.toMatchObject({ id: 'standard' }) + expect(ctx.agentPresets.composedPreset(handle.agent.ctx)).toBe('standard') + expect(warnings).toEqual([expect.stringContaining('tools/change listener failed')]) + stopThrowing() + }) + it('leaves the agent on its previous composition when the new one is unknown', async () => { const handle = await ctx.agents.create({ sessionId: SessionId('sess-unknown'), diff --git a/packages/preset/agent-presets/tsconfig.json b/packages/preset/agent-presets/tsconfig.json index 405ca21e2b..b40cf776ad 100644 --- a/packages/preset/agent-presets/tsconfig.json +++ b/packages/preset/agent-presets/tsconfig.json @@ -30,6 +30,9 @@ { "path": "../../core/system-prompt" }, + { + "path": "../../core/tools" + }, { "path": "../../settings/settings" }, diff --git a/packages/sdk/server/tests/built-scope-carrier.e2e.ts b/packages/sdk/server/tests/built-scope-carrier.e2e.ts index 0aa7099d0b..ae4ab8ef05 100644 --- a/packages/sdk/server/tests/built-scope-carrier.e2e.ts +++ b/packages/sdk/server/tests/built-scope-carrier.e2e.ts @@ -66,7 +66,7 @@ try { const result = Promise.withResolvers(); const unregister = ctx.subagents.registerProvider({ name: "built-local", - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start() { return Promise.resolve({ diff --git a/packages/sdk/server/tests/server.spec.ts b/packages/sdk/server/tests/server.spec.ts index d9527e10d5..257c3e72b0 100644 --- a/packages/sdk/server/tests/server.spec.ts +++ b/packages/sdk/server/tests/server.spec.ts @@ -79,7 +79,7 @@ async function settleSubagent( const result = Promise.withResolvers() const disposeProvider = ctx.subagents.registerProvider({ name: info.provider, - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, async start() { return { @@ -538,7 +538,7 @@ describe('HarnessSdkJsonRpcServer', () => { let currentLocalAgent = oldChild.agent const disposeProvider = ctx.subagents.registerProvider({ name: 'reused', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start() { const result = results[starts] @@ -633,7 +633,7 @@ describe('HarnessSdkJsonRpcServer', () => { const remoteResult = Promise.withResolvers() const unregisterLocal = ctx.subagents.registerProvider({ name: 'reused-provider', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: () => Promise.resolve({ id: SessionId('provider-reuse-child'), @@ -651,7 +651,7 @@ describe('HarnessSdkJsonRpcServer', () => { const unregisterRemote = ctx.subagents.registerProvider({ name: 'reused-provider', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: () => Promise.resolve({ id: SessionId('provider-reuse-child'), @@ -731,7 +731,7 @@ describe('HarnessSdkJsonRpcServer', () => { const missedStartResult = Promise.withResolvers() const disposeMissedStartProvider = ctx.subagents.registerProvider({ name: 'fork', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: true, start: () => Promise.resolve({ id: SessionId('fallback-child-session'), diff --git a/packages/subagent/subagent-acp/README.i18n.yaml b/packages/subagent/subagent-acp/README.i18n.yaml index fa56c715a5..73f55ccfbf 100644 --- a/packages/subagent/subagent-acp/README.i18n.yaml +++ b/packages/subagent/subagent-acp/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/subagent-acp/README.md -README.md: 0201f9bacca5c59031d2fc2eec7fe7210489dd09 -README.zh.md: c987e6251e035f0cf94ac41b570f3be10f0118f5 +README.md: cc4deb5b97152f106caabf747b8c7ccb2f5ddf8e +README.zh.md: e28fb556801d4567bcc606a777e3b090e0551e03 diff --git a/packages/subagent/subagent-acp/README.md b/packages/subagent/subagent-acp/README.md index 0201f9bacc..cc4deb5b97 100644 --- a/packages/subagent/subagent-acp/README.md +++ b/packages/subagent/subagent-acp/README.md @@ -18,7 +18,7 @@ After publication, the provider sends the prompt and collects streamed `agent_me ## Capabilities and context -ACP advertises no start-time capabilities because this process cannot enforce the remote child's depth, tool filter, persona, or structured-output runtime. It also reports `inheritsParentContext: false`: the remote session starts fresh, and the only parent-derived input is the workspace cwd described above — no conversation context crosses the process boundary. +ACP advertises no start-time capabilities because this process cannot apply `request.agentOptions` or enforce the remote child's depth, tool filter, persona, or structured-output runtime. It also reports `inheritsParentContext: false`: the remote session starts fresh, and the only parent-derived input is the workspace cwd described above — no conversation context crosses the process boundary. ## Configuration @@ -70,7 +70,7 @@ The package has no default export. Cordis loader unwrapping would otherwise hide #### What the model sees -The remote child receives the standalone task content through ACP plus its own process's configured system prompt, tools, and fresh session. It receives no parent conversation. This provider advertises no optional start-time capabilities, so the local service rejects requests for persona, tool filtering, depth enforcement, or structured output instead of silently omitting them. +The remote child receives the standalone task content through ACP plus its own process's configured system prompt, tools, and fresh session. It receives no parent conversation. This provider advertises no optional start-time capabilities, so the local service rejects requests for `agentOptions`, persona, tool filtering, depth enforcement, or structured output instead of silently omitting them. #### Token effect @@ -98,6 +98,6 @@ Append-only; newly visible content follows the reusable request prefix and does - **A fresh process per run** — persistent-process pooling is a future optimization ([the seam Agent Note](../../../.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.md)). - **Local workspaces only** — the resolved cwd is a local path handed to a child on the same machine; workspace mapping for a remote ACP agent would need its own backend capability and is not designed here. -- **No optional start-time capabilities** — this provider cannot apply the local harness's `outputSchema`, depth cap, tool filter, or persona inside the remote process, so it advertises none and the service rejects requests that require them. +- **No optional start-time capabilities** — this provider cannot apply the local harness's `agentOptions`, `outputSchema`, depth cap, tool filter, or persona inside the remote process, so it advertises none and the service rejects requests that require them. - **Only committed `agent_message_chunk` text is collected** — the automation server keeps reasoning, tool activity, plans, and other trace data in the child session log rather than emitting them on ACP. - **Permission prompts are auto-answered** (`permission: allow | reject`) — no human is surfaced a child's `session/request_permission`. diff --git a/packages/subagent/subagent-acp/README.zh.md b/packages/subagent/subagent-acp/README.zh.md index c987e6251e..e28fb55680 100644 --- a/packages/subagent/subagent-acp/README.zh.md +++ b/packages/subagent/subagent-acp/README.zh.md @@ -18,7 +18,7 @@ ACP（Agent Client Protocol）提供方会在全新的子进程中运行每个 s ## 能力与上下文 -ACP 不声明任何启动时能力，因为当前进程无法强制执行远程子 agent 的深度、工具过滤、persona 或结构化输出运行时。它也报告 `inheritsParentContext: false`：远程会话从全新状态开始，唯一源自父级的输入是上述工作区 cwd；对话上下文不会跨越进程边界。 +ACP 不声明任何启动时能力，因为当前进程无法应用 `request.agentOptions`，也无法强制执行远程子 agent 的深度、工具过滤、persona 或结构化输出运行时。它也报告 `inheritsParentContext: false`：远程会话从全新状态开始，唯一源自父级的输入是上述工作区 cwd；对话上下文不会跨越进程边界。 ## 配置 @@ -70,7 +70,7 @@ DeepSeek Harness 子进程使用产品启动器和一个显式的绝对路径 `D #### 模型看到的内容 -远程子 agent 通过 ACP 接收独立任务内容，并使用其自身进程配置的系统提示词、工具和全新会话。它不接收父级对话。该提供方不声明任何可选启动时能力，因此本地服务会拒绝要求 persona、工具过滤、深度强制或结构化输出的请求，而不是静默省略这些要求。 +远程子 agent 通过 ACP 接收独立任务内容，并使用其自身进程配置的系统提示词、工具和全新会话。它不接收父级对话。该提供方不声明任何可选启动时能力，因此本地服务会拒绝要求 `agentOptions`、persona、工具过滤、深度强制或结构化输出的请求，而不是静默省略这些要求。 #### Token 影响 @@ -98,6 +98,6 @@ DeepSeek Harness 子进程使用产品启动器和一个显式的绝对路径 `D - **每次运行使用全新进程**：持久进程池属于后续优化（见 [seam Agent Note](../../../.agents/notes/implemented/feature/2026-06-21-subagent-capability-seam.zh.md)）。 - **仅支持本地工作区**：解析后的 cwd 是交给同一台机器上子进程的本地路径；远程 ACP agent 的工作区映射需要独立的后端能力，此处尚未设计这种能力。 -- **不支持可选启动时能力**：该提供方无法在远程进程内应用本地 harness 的 `outputSchema`、深度上限、工具过滤器或 persona，因此不会声明这些能力；服务会拒绝需要它们的请求。 +- **不支持可选启动时能力**：该提供方无法在远程进程内应用本地 harness 的 `agentOptions`、`outputSchema`、深度上限、工具过滤器或 persona，因此不会声明这些能力；服务会拒绝需要它们的请求。 - **只收集已提交的 `agent_message_chunk` 文本**：自动化服务器把推理（reasoning）、工具活动、计划和其他 trace 数据保留在子 agent 会话日志中，不通过 ACP 发出。 - **权限提示自动回答**（`permission: allow | reject`）：不会把子 agent 的 `session/request_permission` 呈现给人。 diff --git a/packages/subagent/subagent-acp/src/index.ts b/packages/subagent/subagent-acp/src/index.ts index 4b526279ba..4c610ef2f9 100644 --- a/packages/subagent/subagent-acp/src/index.ts +++ b/packages/subagent/subagent-acp/src/index.ts @@ -140,11 +140,17 @@ function resolveCwd(configured: string | undefined, request: SubagentStartReques /** * The ACP provider. Advertises NO start-time capabilities: an out-of-process - * child cannot honor `outputSchema`/`maxDepth`/`toolFilter` (the service rejects - * a request needing any of them before `start` runs). + * child cannot honor `agentOptions`/`outputSchema`/`maxDepth`/`toolFilter`/ + * `persona` (the service rejects a request needing any before `start` runs). */ class AcpProvider implements SubagentProvider { - readonly capabilities: SubagentCapabilities = { outputSchema: false, depthLimit: false, toolFilter: false, persona: false } + readonly capabilities: SubagentCapabilities = { + agentOptions: false, + outputSchema: false, + depthLimit: false, + toolFilter: false, + persona: false, + } // Context contract: an out-of-process ACP child starts fresh — no parent conversation crosses the process boundary. readonly inheritsParentContext = false diff --git a/packages/subagent/subagent-acp/tests/subagent-acp.spec.ts b/packages/subagent/subagent-acp/tests/subagent-acp.spec.ts index c534b8e949..d4553d8f86 100644 --- a/packages/subagent/subagent-acp/tests/subagent-acp.spec.ts +++ b/packages/subagent/subagent-acp/tests/subagent-acp.spec.ts @@ -863,7 +863,13 @@ describe('dsh-subagent-acp', () => { it('advertises no start-time capabilities (out-of-process child)', async () => { const ctx = await setup() const provider = ctx.subagents.getProvider('acp')! - expect(provider.capabilities).toEqual({ outputSchema: false, depthLimit: false, toolFilter: false, persona: false }) + expect(provider.capabilities).toEqual({ + agentOptions: false, + outputSchema: false, + depthLimit: false, + toolFilter: false, + persona: false, + }) }) it('unregisters the provider when its fiber is disposed (HMR safety)', async () => { diff --git a/packages/subagent/subagent-claude-code/README.i18n.yaml b/packages/subagent/subagent-claude-code/README.i18n.yaml index ff129c60e2..f894fb15ab 100644 --- a/packages/subagent/subagent-claude-code/README.i18n.yaml +++ b/packages/subagent/subagent-claude-code/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/subagent-claude-code/README.md -README.md: 0260d9c82dee82541e5b60ff7fe2c323cf9b7331 -README.zh.md: 484241f1ca23f2b8ee843b6f412d0f779832d400 +README.md: 30b61347dfda110fe95a9153aeea35c8f760012f +README.zh.md: 17ddac7d4beebfbf0f374f606a68d77b13ec2b77 diff --git a/packages/subagent/subagent-claude-code/README.md b/packages/subagent/subagent-claude-code/README.md index 0260d9c82d..30b61347df 100644 --- a/packages/subagent/subagent-claude-code/README.md +++ b/packages/subagent/subagent-claude-code/README.md @@ -20,7 +20,7 @@ Each query sets `persistSession: false` and disables `AskUserQuestion`. Except i ## Capabilities and context -The provider advertises no optional start-time capabilities and reports `inheritsParentContext: false`. Claude Code receives the standalone text task and the parent Session cwd, but not the parent conversation, persona, tool filter, depth policy, or structured-output contract. Every run has an independent SDK query, cancellation controller, CLI process, and non-persisted product session. +The provider advertises no optional start-time capabilities and reports `inheritsParentContext: false`. The shared service rejects `request.agentOptions` for this provider. Claude Code receives the standalone text task and the parent Session cwd, but not the parent conversation, persona, tool filter, depth policy, or structured-output contract. Every run has an independent SDK query, cancellation controller, CLI process, and non-persisted product session. ## Configuration @@ -145,5 +145,5 @@ Append-only: foreground adds one result after the reusable parent prefix, while - **The SDK platform payload is required at delegation time** — installs that omit optional dependencies, unsupported platforms, and missing or damaged payloads fail at the first query; there is no host-CLI fallback. - **No human interaction path** — `AskUserQuestion` is disabled, permission prompts are denied, MCP elicitation is declined, and blocking dialogs fail closed instead of suspending. - **Assistant payload is final text only** — a failed run may additionally expose the separate safe diagnostic; reasoning, intermediate messages, tool traffic, usage, stderr, and workspace diffs remain product-local, while generic Job ids, notices, and status come from the shared job runtime. -- **No optional shared capabilities** — output schemas, child personas, tool filtering, and harness depth enforcement are rejected by the shared service for this provider. +- **No optional shared capabilities** — `agentOptions`, output schemas, child personas, tool filtering, and harness depth enforcement are rejected by the shared service for this provider. - **No wall-clock timeout or side-effect rollback** — the caller cancels long work, and files or external systems changed before cancellation are not restored. diff --git a/packages/subagent/subagent-claude-code/README.zh.md b/packages/subagent/subagent-claude-code/README.zh.md index 484241f1ca..17ddac7d4b 100644 --- a/packages/subagent/subagent-claude-code/README.zh.md +++ b/packages/subagent/subagent-claude-code/README.zh.md @@ -20,7 +20,7 @@ SDK 接收由文本块原样拼接成的任务。提供方会完整迭代 SDK ## 能力与上下文 -本提供方不声明任何可选的启动时能力，并报告 `inheritsParentContext: false`。Claude Code 会接收独立文本任务和父会话 cwd，但不会接收父会话的对话、角色设定、工具筛选器、深度策略或结构化输出约定。每次运行都拥有独立的 SDK query、取消控制器、CLI 进程和不持久化的产品会话。 +本提供方不声明任何可选的启动时能力，并报告 `inheritsParentContext: false`。共享服务会拒绝本提供方的 `request.agentOptions`。Claude Code 会接收独立文本任务和父会话 cwd，但不会接收父会话的对话、角色设定、工具筛选器、深度策略或结构化输出约定。每次运行都拥有独立的 SDK query、取消控制器、CLI 进程和不持久化的产品会话。 ## 配置 @@ -145,5 +145,5 @@ Claude Code 子级会在一个全新的 SDK query 中接收独立文本任务。 - **委派时必须存在 SDK 平台载荷**：省略 optional dependencies 的安装、不受支持的平台以及缺失或损坏的载荷都会在第一次 query 时失败；不会回退到宿主 CLI。 - **没有人工交互路径**：`AskUserQuestion` 被禁用，权限提示会被拒绝，MCP elicitation 会被拒绝，阻塞对话会快速失败而不会挂起。 - **assistant 载荷仅包含最终文本**：失败运行可以额外公开独立的安全诊断；推理、中间消息、工具通信、用量信息、stderr 和工作区差异仍只保留在产品内部，通用 Job id、通知与状态来自共享作业运行时。 -- **没有可选的共享能力**：对于本提供方，共享服务会拒绝输出 schema、子任务角色设定、工具筛选和 harness 深度强制约束。 +- **没有可选的共享能力**：对于本提供方，共享服务会拒绝 `agentOptions`、输出 schema、子任务角色设定、工具筛选和 harness 深度强制约束。 - **没有按实际经过时间触发的超时或副作用回滚**：长时间运行的工作由调用方取消，且取消前已更改的文件或外部系统不会恢复原状。 diff --git a/packages/subagent/subagent-claude-code/tests/loader-composition.e2e.ts b/packages/subagent/subagent-claude-code/tests/loader-composition.e2e.ts index 5f339ce22e..6054370615 100644 --- a/packages/subagent/subagent-claude-code/tests/loader-composition.e2e.ts +++ b/packages/subagent/subagent-claude-code/tests/loader-composition.e2e.ts @@ -45,6 +45,7 @@ describe('product-provider public Loader composition', () => { { name: 'codex', capabilities: { + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, @@ -55,6 +56,7 @@ describe('product-provider public Loader composition', () => { { name: 'claude-code', capabilities: { + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, @@ -65,6 +67,7 @@ describe('product-provider public Loader composition', () => { { name: 'claude-primary', capabilities: { + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, @@ -75,6 +78,7 @@ describe('product-provider public Loader composition', () => { { name: 'claude-secondary', capabilities: { + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, diff --git a/packages/subagent/subagent-codex/README.i18n.yaml b/packages/subagent/subagent-codex/README.i18n.yaml index 0f8ed31eab..fd0cccc02d 100644 --- a/packages/subagent/subagent-codex/README.i18n.yaml +++ b/packages/subagent/subagent-codex/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/subagent-codex/README.md -README.md: 975f353b9f1bc6fab61a4c0eb40ebaf50c436623 -README.zh.md: 2ea256afb3bb8fdfe57fd555b6db10522a709a56 +README.md: 5016d6b9aa4b57c9d1e83701a0ccfc4b04616f6d +README.zh.md: adabd942c362a28a4eeabe7a4d1b27b138aee4c6 diff --git a/packages/subagent/subagent-codex/README.md b/packages/subagent/subagent-codex/README.md index 975f353b9f..5016d6b9aa 100644 --- a/packages/subagent/subagent-codex/README.md +++ b/packages/subagent/subagent-codex/README.md @@ -18,7 +18,7 @@ Local cancellation wins the result race and maps to `aborted`. For failed turns, ## Capabilities and context -The provider advertises no optional start-time capabilities and reports `inheritsParentContext: false`. Codex receives the standalone text task and the parent Session cwd, but not the parent conversation, persona, tool filter, depth policy, or structured-output contract. The ephemeral Codex thread id and turn id stay private to this run and are never persisted in the parent Session. +The provider advertises no optional start-time capabilities and reports `inheritsParentContext: false`. The shared service rejects `request.agentOptions` for this provider. Codex receives the standalone text task and the parent Session cwd, but not the parent conversation, persona, tool filter, depth policy, or structured-output contract. The ephemeral Codex thread id and turn id stay private to this run and are never persisted in the parent Session. ## Configuration @@ -139,5 +139,5 @@ Append-only: foreground adds one result after the reusable parent prefix, while - **Compatibility is pinned by development evidence** — upgrading from the verified 0.147.0 protocol baseline requires regenerating upstream schema evidence and rerunning handshake, answer-selection, approval, cancellation, keyless real-product, and credentialed DeepSeek nonce tests. - **No human approval path** — known unattended approval requests are denied and unknown server requests fail closed; the three Profile modes never create a DSH interaction channel or per-call allow policy. - **Assistant payload is final text only** — a failed run may additionally expose the separate safe diagnostic; reasoning, commentary, intermediate messages, tool traffic, usage, raw stderr, and workspace diffs remain outside the parent Session, while generic Job ids, notices, and status come from the shared job runtime. -- **No optional shared capabilities** — output schemas, child personas, tool filtering, and harness depth enforcement are rejected by the shared service for this provider. +- **No optional shared capabilities** — `agentOptions`, output schemas, child personas, tool filtering, and harness depth enforcement are rejected by the shared service for this provider. - **No wall-clock timeout or side-effect rollback** — the caller cancels long work, and files or external systems changed before cancellation are not restored. diff --git a/packages/subagent/subagent-codex/README.zh.md b/packages/subagent/subagent-codex/README.zh.md index 2ea256afb3..adabd942c3 100644 --- a/packages/subagent/subagent-codex/README.zh.md +++ b/packages/subagent/subagent-codex/README.zh.md @@ -18,7 +18,7 @@ ## 能力与上下文 -本提供方不声明任何可选的启动时能力，并报告 `inheritsParentContext: false`。Codex 会接收独立文本任务和父会话 cwd，但不会接收父会话的对话、角色设定、工具筛选器、深度策略或结构化输出约定。临时 Codex 线程 ID 与轮次 ID 仅在此次运行内部可见，绝不会持久化到父会话。 +本提供方不声明任何可选的启动时能力，并报告 `inheritsParentContext: false`。共享服务会拒绝本提供方的 `request.agentOptions`。Codex 会接收独立文本任务和父会话 cwd，但不会接收父会话的对话、角色设定、工具筛选器、深度策略或结构化输出约定。临时 Codex 线程 ID 与轮次 ID 仅在此次运行内部可见，绝不会持久化到父会话。 ## 配置 @@ -139,5 +139,5 @@ Codex 子级会在一个全新的临时线程中，以单个轮次接收这些 - **兼容性由开发证据锁定**：若要从已验证的 0.147.0 协议基线升级，必须重新生成上游 schema 证据，并重新运行握手、答案选择、审批、取消、无密钥真实产品以及带密钥的 DeepSeek 随机数测试。 - **没有人工审批路径**：已知的无人值守审批请求会被拒绝，未知服务器请求会以默认拒绝方式使运行失败；三种 Profile 模式都不会创建 DSH 交互通道或逐次调用 allow 策略。 - **assistant 载荷仅包含最终文本**：失败运行可以额外公开独立的安全诊断；推理、过程说明、中间消息、工具通信、用量信息、原始 stderr 和工作区差异不会进入父会话，通用 Job id、通知与状态来自共享作业运行时。 -- **没有可选的共享能力**：对于本提供方，共享服务会拒绝输出 schema、子任务角色设定、工具筛选和 harness 深度强制约束。 +- **没有可选的共享能力**：对于本提供方，共享服务会拒绝 `agentOptions`、输出 schema、子任务角色设定、工具筛选和 harness 深度强制约束。 - **没有按实际经过时间触发的超时或副作用回滚**：长时间运行的工作由调用方取消，且取消前已更改的文件或外部系统不会恢复原状。 diff --git a/packages/subagent/subagent-codex/tests/loader-composition.e2e.ts b/packages/subagent/subagent-codex/tests/loader-composition.e2e.ts index b55b819989..c411ae2dbf 100644 --- a/packages/subagent/subagent-codex/tests/loader-composition.e2e.ts +++ b/packages/subagent/subagent-codex/tests/loader-composition.e2e.ts @@ -45,6 +45,7 @@ describe('Codex provider public Loader composition', () => { { name: 'codex', capabilities: { + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, @@ -55,6 +56,7 @@ describe('Codex provider public Loader composition', () => { { name: 'codex-primary', capabilities: { + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, @@ -65,6 +67,7 @@ describe('Codex provider public Loader composition', () => { { name: 'codex-secondary', capabilities: { + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, diff --git a/packages/subagent/subagent-dsh-sdk/README.i18n.yaml b/packages/subagent/subagent-dsh-sdk/README.i18n.yaml index 8946de4236..a05ec421e3 100644 --- a/packages/subagent/subagent-dsh-sdk/README.i18n.yaml +++ b/packages/subagent/subagent-dsh-sdk/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/subagent-dsh-sdk/README.md -README.md: 10f437f1618d006f58ad37701e6fde7944f5c1fb -README.zh.md: d3237d9a0b36ce63c9eacf85bfe7cff947acd9e0 +README.md: 302baa05afed2b78c2041f57b1ef900713e350cd +README.zh.md: e4e1460274170b0c990d9e454f1cbf54546676e9 diff --git a/packages/subagent/subagent-dsh-sdk/README.md b/packages/subagent/subagent-dsh-sdk/README.md index 10f437f161..302baa05af 100644 --- a/packages/subagent/subagent-dsh-sdk/README.md +++ b/packages/subagent/subagent-dsh-sdk/README.md @@ -20,7 +20,7 @@ The SDK client returns an owned child activity rather than a prompt result. The ## Capabilities and context -The provider advertises no start-time capabilities (`outputSchema`/`depthLimit`/`toolFilter`/`persona` all false) and `inheritsParentContext: false`: the child is a fresh runtime in another process, and the only parent-derived input is the workspace cwd. `dsh-tool-subagent` deployments over this provider set `maxDepth: 'provider-managed'` — the child harness owns its own recursion budget. +The provider advertises no start-time capabilities (`agentOptions`/`outputSchema`/`depthLimit`/`toolFilter`/`persona` all false) and `inheritsParentContext: false`: the child is a fresh runtime in another process, and the only parent-derived input is the workspace cwd. `dsh-tool-subagent` deployments over this provider set `maxDepth: 'provider-managed'` — the child harness owns its own recursion budget. ## Configuration @@ -68,7 +68,7 @@ The package has no default export. Cordis loader unwrapping would otherwise hide #### What the model sees -The child runtime's model receives the standalone task as its user message plus that runtime's own configured system prompt, tools, and fresh session. It receives no parent conversation. This provider advertises no optional start-time capabilities, so the local service rejects requests for persona, tool filtering, depth enforcement, or structured output instead of silently omitting them. +The child runtime's model receives the standalone task as its user message plus that runtime's own configured system prompt, tools, and fresh session. It receives no parent conversation. This provider advertises no optional start-time capabilities, so the local service rejects requests for `agentOptions`, persona, tool filtering, depth enforcement, or structured output instead of silently omitting them. #### Token effect @@ -95,6 +95,6 @@ Append-only; newly visible content follows the reusable request prefix and does ## Known Limitations and Deferred Work - **A fresh runtime process per run** — no pooling; a harness runtime boots a full plugin tree, so per-run spawn cost is higher than the ACP backend's typical child. -- **No optional start-time capabilities** — the parent cannot enforce `outputSchema`, depth, tool filters, or persona inside the child process; configure the selected child profile and its ordered patches instead. +- **No optional start-time capabilities** — the parent cannot apply `agentOptions` or enforce `outputSchema`, depth, tool filters, or persona inside the child process; configure the selected child profile and its ordered patches instead. - **The child's transcript stays in the child's own session root** — the parent log records only the delegation tool call/result (the seam's child-isolation rule); the streamed `session.event` channel is consumed for output extraction, not bridged into the parent log. - **Local child processes only** — the resolved cwd is a local path; a remote runtime would need its own backend. diff --git a/packages/subagent/subagent-dsh-sdk/README.zh.md b/packages/subagent/subagent-dsh-sdk/README.zh.md index d3237d9a0b..e4e1460274 100644 --- a/packages/subagent/subagent-dsh-sdk/README.zh.md +++ b/packages/subagent/subagent-dsh-sdk/README.zh.md @@ -20,7 +20,7 @@ SDK 客户端返回自有子活动，而不是提示词结果。提供方读取 ## 能力与上下文 -Provider 不宣告任何启动期能力（`outputSchema`/`depthLimit`/`toolFilter`/`persona` 全为 false），且 `inheritsParentContext: false`：子进程是另一进程里的全新运行时，唯一来自父方的输入是工作区 cwd。基于本 provider 的 `dsh-tool-subagent` 部署应设置 `maxDepth: 'provider-managed'`——子 harness 拥有自己的递归预算。 +Provider 不宣告任何启动期能力（`agentOptions`/`outputSchema`/`depthLimit`/`toolFilter`/`persona` 全为 false），且 `inheritsParentContext: false`：子进程是另一进程里的全新运行时，唯一来自父方的输入是工作区 cwd。基于本 provider 的 `dsh-tool-subagent` 部署应设置 `maxDepth: 'provider-managed'`——子 harness 拥有自己的递归预算。 ## 配置 @@ -68,7 +68,7 @@ Provider 不宣告任何启动期能力（`outputSchema`/`depthLimit`/`toolFilte #### 模型看到的内容 -子运行时的模型会收到作为用户消息的独立任务，以及该运行时自身配置的系统提示词、工具和全新会话。它不会收到父级对话。本提供方不声明可选的启动时能力，因此本地服务会拒绝要求 persona、工具过滤、深度强制或结构化输出的请求，而不是静默省略这些要求。 +子运行时的模型会收到作为用户消息的独立任务，以及该运行时自身配置的系统提示词、工具和全新会话。它不会收到父级对话。本提供方不声明可选的启动时能力，因此本地服务会拒绝要求 `agentOptions`、persona、工具过滤、深度强制或结构化输出的请求，而不是静默省略这些要求。 #### Token 影响 @@ -95,6 +95,6 @@ Provider 不宣告任何启动期能力（`outputSchema`/`depthLimit`/`toolFilte ## 已知限制与暂缓事项 - **每次运行都使用全新的运行时进程**：不使用进程池；harness 运行时需要启动完整的插件树，因此每次运行的 spawn 成本高于 ACP 后端通常使用的子进程。 -- **不支持可选的启动时能力**：父级无法在子进程内强制执行 `outputSchema`、深度限制、工具过滤或 persona；应改为配置所选子 profile 及其有序 patch。 +- **不支持可选的启动时能力**：父级无法在子进程内应用 `agentOptions`，也无法强制执行 `outputSchema`、深度限制、工具过滤或 persona；应改为配置所选子 profile 及其有序 patch。 - **子进程的 transcript（文本记录）保留在其自身的会话根目录中**：父级日志只记录委派工具调用／结果（seam 的子级隔离规则）；流式 `session.event` 通道只用于提取输出，不会桥接到父级日志中。 - **仅支持本地子进程**：解析出的 cwd 是本地路径；远程运行时需要独立的后端。 diff --git a/packages/subagent/subagent-dsh-sdk/src/index.ts b/packages/subagent/subagent-dsh-sdk/src/index.ts index 7530e104d2..277add74f8 100644 --- a/packages/subagent/subagent-dsh-sdk/src/index.ts +++ b/packages/subagent/subagent-dsh-sdk/src/index.ts @@ -105,7 +105,7 @@ function resolveConfiguredFile(field: string, value: string): string { /** * The SDK provider. Advertises NO start-time capabilities: an out-of-process - * child cannot honor `outputSchema`/`maxDepth`/`toolFilter`/`persona` (the + * child cannot honor `agentOptions`/`outputSchema`/`maxDepth`/`toolFilter`/`persona` (the * service rejects a request needing any of them before `start` runs). */ class SdkSubagentProvider implements SubagentProvider { diff --git a/packages/subagent/subagent-dsh-sdk/tests/subagent-dsh-sdk.spec.ts b/packages/subagent/subagent-dsh-sdk/tests/subagent-dsh-sdk.spec.ts index 2e67c76c24..27511670cf 100644 --- a/packages/subagent/subagent-dsh-sdk/tests/subagent-dsh-sdk.spec.ts +++ b/packages/subagent/subagent-dsh-sdk/tests/subagent-dsh-sdk.spec.ts @@ -433,6 +433,7 @@ describe('dsh-subagent-dsh-sdk provider', () => { expect(ctx.subagents.getProvider('sdk-hmr')?.name).toBe('sdk-hmr') expect(ctx.subagents.getProvider('sdk-hmr')?.inheritsParentContext).toBe(false) expect(ctx.subagents.getProvider('sdk-hmr')?.capabilities).toEqual({ + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, diff --git a/packages/subagent/subagent-fork-in-process/README.i18n.yaml b/packages/subagent/subagent-fork-in-process/README.i18n.yaml index 9cd2361d52..19af9e598a 100644 --- a/packages/subagent/subagent-fork-in-process/README.i18n.yaml +++ b/packages/subagent/subagent-fork-in-process/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/subagent-fork-in-process/README.md -README.md: 74c27ff10c76aa711ed3e954e806c00a27aacfa5 -README.zh.md: 43e7ef489b33d52b674420d08f7edf8c89fb0e42 +README.md: c2dcda39c03b8c059839bef1573436e485a51911 +README.zh.md: 3eb84053a3e74abe7394c0985dcd88bd6146e449 diff --git a/packages/subagent/subagent-fork-in-process/README.md b/packages/subagent/subagent-fork-in-process/README.md index 74c27ff10c..c2dcda39c0 100644 --- a/packages/subagent/subagent-fork-in-process/README.md +++ b/packages/subagent/subagent-fork-in-process/README.md @@ -16,7 +16,7 @@ The seed transfers conversation history only. The child still receives a fresh f `start(request)` passes the completed-turn seed to [`startInProcessRun`](../subagent-in-process-driver/README.md) and awaits child publication. The shared driver owns cancellation, depth, customization, result reading, and disposal. -Fork advertises `{ outputSchema: true, depthLimit: true, toolFilter: true, persona: true }`, identical to spawn. +Fork advertises `{ agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }`, identical to spawn. ## Config @@ -39,7 +39,7 @@ Forking duplicates retained completed history into separate child requests; the #### KV Cache effect -The child may reuse the inherited byte-identical prefix under the same provider and model. Persona, tool-filter, generated-SDK, or route changes may invalidate reuse before inherited history; later child history is append-only. Shipped compositions therefore bind this provider to `backgroundMode: one-shot`, because a continuable child additionally carries the child-scoped `report` tool and its prompt section — deltas that precede the inherited history and so invalidate all of it ([the fork-one-shot Agent Note](../../../.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md)). +The child may reuse the inherited byte-identical prefix under the same provider and model. Persona, tool-filter, generated-SDK, or route changes may invalidate reuse before inherited history; later child history is append-only. The base bundle and ACP/headless examples bind this provider to `backgroundMode: one-shot`, because a continuable child additionally carries the child-scoped `report` tool and its prompt section — deltas that precede the inherited history and so invalidate all of it. The CLI presets retain `continuable` fork and therefore accept that prefix loss ([the cache-preserving fork Agent Note](../../../.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md)). ### Parent tool result, indirectly @@ -58,4 +58,5 @@ Append-only; newly visible content follows the reusable request prefix and does ## Known Limitations and Deferred Work - **The seed is a one-time snapshot** — the child sees the parent's completed turns as of the fork and nothing the parent logs afterwards; there is no live context sharing. -- **No shipped composition creates a continuable fork child** — `prepareContinuable` remains implemented and the seam accepts it, but every shipped `cordis.yml` sets `backgroundMode: one-shot` on the fork delegation tool, so the provider's continuable path has no production caller. Reopening it requires the child's system prompt and tool schemas to match the parent's byte for byte, which the [`report` return channel](../tool-subagent-report/README.md) currently prevents. Rationale and the reintroduction condition: [the fork-one-shot Agent Note](../../../.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md). +- **Fork lifecycle policy differs by composition** — the base bundle and ACP/headless examples use one-shot fork to preserve prefix reuse, while the CLI presets use continuable fork and accept the child-scoped [`report` return channel](../tool-subagent-report/README.md) invalidating that prefix. Making continuable fork cache-preserving requires the child system prompt and tool schemas to match the parent's byte for byte. Rationale and the reintroduction condition: [the cache-preserving fork Agent Note](../../../.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md). +- **Shipped fork tools do not expose child LLM route selection** — they inherit the parent's provider and model so the copied history remains eligible for KV Cache reuse. Route selection stays disabled until a change can preserve reuse or expose a bounded recomputation cost; the [model-selected route Agent Note](../../../.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md) owns that separate restriction. diff --git a/packages/subagent/subagent-fork-in-process/README.zh.md b/packages/subagent/subagent-fork-in-process/README.zh.md index 43e7ef489b..3eb84053a3 100644 --- a/packages/subagent/subagent-fork-in-process/README.zh.md +++ b/packages/subagent/subagent-fork-in-process/README.zh.md @@ -16,7 +16,7 @@ subagent 启动时，父 agent 当前的工具调用轮次仍未结束：其日 `start(request)` 将已完成轮次的初始内容传给 [`startInProcessRun`](../subagent-in-process-driver/README.zh.md)，并等待子 agent 发布。共享驱动器负责取消、深度、定制、结果读取和 dispose（资源释放）。 -fork 声明 `{ outputSchema: true, depthLimit: true, toolFilter: true, persona: true }`，与 spawn 相同。 +fork 声明 `{ agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }`，与 spawn 相同。 ## 配置 @@ -39,7 +39,7 @@ fork 会把保留的已完成历史复制到独立的子 agent 请求中；随 #### KV Cache 影响 -在提供方和模型相同的前提下，子 agent 可以复用继承的逐字节相同前缀。persona、工具过滤、生成 SDK 或路由变化可能在继承历史之前使复用失效；后续子 agent 历史仅追加。因此随附组合把本提供方绑定为 `backgroundMode: one-shot`：可继续子 agent 还会额外携带作用域局部的 `report` 工具及其提示词 section，而这些增量位于继承历史之前，会使继承历史整体失效（见 [fork 保持 one-shot 的 Agent Note](../../../.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.zh.md)）。 +在提供方和模型相同的前提下，子 agent 可以复用继承的逐字节相同前缀。persona、工具过滤、生成 SDK 或路由变化可能在继承历史之前使复用失效；后续子 agent 历史仅追加。base 组合包与 ACP/headless 示例把本提供方绑定为 `backgroundMode: one-shot`：可继续子 agent 还会额外携带作用域局部的 `report` 工具及其提示词 section，而这些增量位于继承历史之前，会使继承历史整体失效。CLI preset 保留可继续 fork，因此接受这项前缀损失（见[保留缓存的 fork Agent Note](../../../.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.zh.md)）。 ### 父 agent 工具结果（间接） @@ -58,4 +58,5 @@ fork 会把保留的已完成历史复制到独立的子 agent 请求中；随 ## 已知限制与暂缓事项 - **初始内容是一次性快照**：子 agent 只能看到 fork 时父 agent 已完成的轮次，看不到父 agent 此后记录的任何内容；不会实时共享上下文。 -- **没有任何随附组合会创建可继续的 fork 子 agent**：`prepareContinuable` 仍然实现完好，seam 也接受它，但每份随附的 `cordis.yml` 都在 fork 委派工具上设置 `backgroundMode: one-shot`，因此该提供方的可继续路径没有生产调用方。重新开放它需要子 agent 的系统提示词与工具 schema 与父 agent 逐字节一致，而这一点目前被 [`report` 返回通道](../tool-subagent-report/README.zh.md)阻止。理由与重新开放条件见 [fork 保持 one-shot 的 Agent Note](../../../.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.zh.md)。 +- **fork 生命周期策略因组合而异**：base 组合包与 ACP/headless 示例使用一次性 fork 以保留前缀复用，CLI preset 则使用可继续 fork，并接受子级作用域的 [`report` 返回通道](../tool-subagent-report/README.zh.md)使该前缀失效。要让可继续 fork 保留缓存，子 agent 的系统提示词与工具 schema 必须与父级逐字节一致。理由与重新开放条件见[保留缓存的 fork Agent Note](../../../.agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.zh.md)。 +- **随附 fork 工具不公开子级 LLM 路由选择**：它们会继承父级的提供方与模型，使复制的历史仍可供 KV Cache 复用。只有在路由变化仍能保留复用，或接口能公开一项有界的重算成本时，才启用路由选择；该独立限制由[模型选择路由 Agent Note](../../../.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.zh.md)负责。 diff --git a/packages/subagent/subagent-fork-in-process/src/index.ts b/packages/subagent/subagent-fork-in-process/src/index.ts index 1f8e48b8c8..9786f585fa 100644 --- a/packages/subagent/subagent-fork-in-process/src/index.ts +++ b/packages/subagent/subagent-fork-in-process/src/index.ts @@ -55,11 +55,18 @@ function completedTurnPrefix(parent: Agent): SessionEvent[] { /** * The fork provider. Supports `depthLimit` and `outputSchema` (via the shared - * in-process structured runtime), plus `toolFilter`/`persona` (scoped - * restrict() and a scoped shadowing persona section). + * in-process structured runtime), `agentOptions` (merged over the parent + * route), and `toolFilter`/`persona` (scoped restrict() and a scoped shadowing + * persona section). */ class ForkInProcessProvider implements SubagentProvider { - readonly capabilities: SubagentCapabilities = { outputSchema: true, depthLimit: true, toolFilter: true, persona: true } + readonly capabilities: SubagentCapabilities = { + agentOptions: true, + outputSchema: true, + depthLimit: true, + toolFilter: true, + persona: true, + } // Context contract: a forked child IS seeded with the parent's completed-turn prefix. readonly inheritsParentContext = true @@ -74,11 +81,11 @@ class ForkInProcessProvider implements SubagentProvider { }) } - // TODO(fork-continuable-prefix-reuse): no shipped composition calls this — - // they bind fork to `backgroundMode: one-shot` because a continuable child's - // `report` tool and prompt section precede the inherited history, defeating - // the prefix reuse a fork exists for. Reopening needs a byte-identical child - // system prompt and tool schemas; see issue #2124 and + // TODO(fork-continuable-prefix-reuse): CLI presets call this and accept that + // a continuable child's `report` tool and prompt section precede the inherited + // history, defeating the prefix reuse a fork exists for. Cache-preserving + // continuable fork needs byte-identical child system prompt and tool schemas; + // see issue #2124 and // .agents/notes/implemented/architecture/2026-08-10-fork-children-stay-one-shot.md. prepareContinuable(request: ContinuableCreateRequest): Promise { // The fork prefix is captured ONCE, at creation: it becomes part of the diff --git a/packages/subagent/subagent-fork-in-process/tests/subagent-fork-in-process.spec.ts b/packages/subagent/subagent-fork-in-process/tests/subagent-fork-in-process.spec.ts index 8b69711233..292fba32cc 100644 --- a/packages/subagent/subagent-fork-in-process/tests/subagent-fork-in-process.spec.ts +++ b/packages/subagent/subagent-fork-in-process/tests/subagent-fork-in-process.spec.ts @@ -193,9 +193,15 @@ describe('dsh-subagent-fork-in-process', () => { await run.dispose() }) - it('advertises every start-time capability (depthLimit, outputSchema, toolFilter, persona)', async () => { + it('advertises every start-time capability', async () => { const { ctx } = await setup([]) - expect(ctx.subagents.getProvider('fork')!.capabilities).toEqual({ outputSchema: true, depthLimit: true, toolFilter: true, persona: true }) + expect(ctx.subagents.getProvider('fork')!.capabilities).toEqual({ + agentOptions: true, + outputSchema: true, + depthLimit: true, + toolFilter: true, + persona: true, + }) }) it('unregisters the provider when its fiber is disposed (HMR safety)', async () => { diff --git a/packages/subagent/subagent-in-process-driver/README.i18n.yaml b/packages/subagent/subagent-in-process-driver/README.i18n.yaml index 505c480cfe..16d07d8211 100644 --- a/packages/subagent/subagent-in-process-driver/README.i18n.yaml +++ b/packages/subagent/subagent-in-process-driver/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/subagent-in-process-driver/README.md -README.md: 47a5c09fc1c80c5dc3062be82e7355b874a627d3 -README.zh.md: b96399795a0fbac05ef1795888aa93c620687f30 +README.md: ed2568fcff3fe1f0f3968d1cef43ebd914a8911b +README.zh.md: f9958e5c2b819d51bfdf8fc1e14d1f8c7c19be91 diff --git a/packages/subagent/subagent-in-process-driver/README.md b/packages/subagent/subagent-in-process-driver/README.md index 47a5c09fc1..ed2568fcff 100644 --- a/packages/subagent/subagent-in-process-driver/README.md +++ b/packages/subagent/subagent-in-process-driver/README.md @@ -16,7 +16,7 @@ The driver follows this sequence: 4. Publish the child, retain the returned `AgentHandle`, and drive one task with `child.followup(prompt)` followed by `child.whenIdle()`. 5. Read the child's own output — its last non-empty assistant message (an empty-content message that records usage is skipped), or its accumulated assistant text when no such message exists — and the final durable turn reason from the complete owned child run, excluding any fork seed. -The child gets the parent's working-directory/session lineage and inherits the parent provider, model, and output-token cap unless `request.agentOptions` overrides them. It gets a fresh flat registration scope: parent ownership does not import parent tool restrictions or establish an authority subset. +The child gets the parent's working-directory/session lineage and inherits the parent provider, model, reasoning effort, and output-token cap unless `request.agentOptions` overrides them. It gets a fresh flat registration scope: parent ownership does not import parent tool restrictions or establish an authority subset. This result boundary is valid because the provider owns an isolated child lifecycle from publication through quiescence. Steering submitted during that lifecycle belongs to the child run; the provider does not pretend the initial follow-up alone owns its output. diff --git a/packages/subagent/subagent-in-process-driver/README.zh.md b/packages/subagent/subagent-in-process-driver/README.zh.md index b96399795a..f9958e5c2b 100644 --- a/packages/subagent/subagent-in-process-driver/README.zh.md +++ b/packages/subagent/subagent-in-process-driver/README.zh.md @@ -16,7 +16,7 @@ 4. 发布子 agent，保留返回的 `AgentHandle`，并通过先调用 `child.followup(prompt)`、再调用 `child.whenIdle()` 来驱动一项任务。 5. 从完整的自有子运行中读取子 agent 自身的输出——最后一条非空 assistant 消息（记录 usage 的空内容消息会被跳过），若没有这类消息则取其累积的 assistant 文本——以及最终持久化的轮次原因，并排除任何 fork 初始内容。 -子 agent 会获得父 agent 的工作目录／会话谱系；除非 `request.agentOptions` 覆盖，否则还会继承父 agent 的提供方、模型和输出 token 上限。它获得全新的扁平注册作用域：父级所有权不会导入父 agent 的工具限制，也不会建立权限子集。 +子 agent 会获得父 agent 的工作目录／会话谱系；除非 `request.agentOptions` 覆盖，否则还会继承父 agent 的提供方、模型、推理强度与输出 token 上限。它获得全新的扁平注册作用域：父级所有权不会导入父 agent 的工具限制，也不会建立权限子集。 该结果边界成立，是因为提供方拥有从发布到完全停稳的隔离子 agent 生命周期。在该生命周期内提交的 steering（中途引导）属于子运行；提供方不会声称输出只归初始 follow-up 所有。 diff --git a/packages/subagent/subagent-in-process-driver/tests/structured.spec.ts b/packages/subagent/subagent-in-process-driver/tests/structured.spec.ts index 8840e8b101..6bb18f5f28 100644 --- a/packages/subagent/subagent-in-process-driver/tests/structured.spec.ts +++ b/packages/subagent/subagent-in-process-driver/tests/structured.spec.ts @@ -70,7 +70,7 @@ async function setup(script: Script, options: SetupOptions = {}) { await ctx.plugin(SubagentRuntime) const disposeProvider = ctx.subagents.registerProvider({ name: 'spawn', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: false, persona: false }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: false, persona: false }, inheritsParentContext: false, start: (request: ResolvedSubagentStartRequest) => startInProcessRun(request, {}), }) diff --git a/packages/subagent/subagent-spawn-in-process/README.i18n.yaml b/packages/subagent/subagent-spawn-in-process/README.i18n.yaml index 1246456cc0..022dbfa58d 100644 --- a/packages/subagent/subagent-spawn-in-process/README.i18n.yaml +++ b/packages/subagent/subagent-spawn-in-process/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/subagent-spawn-in-process/README.md -README.md: f1fb96f2230359cb3ff55c630f29fd34345dbed7 -README.zh.md: 95a3b5cdb7084eb75666f8d62001221c57ac676c +README.md: ebe2b069dc56dc1a3359f8880860a5796ef3ef4c +README.zh.md: 66ecbec2c00865d16a99f1e6bf4f0c32cb6e5538 diff --git a/packages/subagent/subagent-spawn-in-process/README.md b/packages/subagent/subagent-spawn-in-process/README.md index f1fb96f223..ebe2b069dc 100644 --- a/packages/subagent/subagent-spawn-in-process/README.md +++ b/packages/subagent/subagent-spawn-in-process/README.md @@ -6,13 +6,13 @@ The spawn provider creates a fresh child `Agent` in the current process. The chi ## Behavior -`start(request)` delegates to [`startInProcessRun`](../subagent-in-process-driver/README.md) with no seed and awaits publication before returning. The child receives parent working-directory/session lineage and inherits the parent model unless overridden, but starts with an empty conversation. +`start(request)` delegates to [`startInProcessRun`](../subagent-in-process-driver/README.md) with no seed and awaits publication before returning. The child receives parent working-directory/session lineage and inherits the parent provider, model, reasoning effort, and output-token limit unless `request.agentOptions` overrides them, but starts with an empty conversation. The shared driver owns depth checking, persona and tool-filter setup, structured output, required-signal cancellation, one-shot execution, result reading, and quiescent disposal. A startup rejection leaves no published child; provider unload after fulfillment does not revoke the holder-owned run. ## Capabilities -Spawn advertises `{ outputSchema: true, depthLimit: true, toolFilter: true, persona: true }` because it controls the child's creation window and can enforce all four features. +Spawn advertises `{ agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }` because it controls the child's creation window and can enforce all five features. ## Config @@ -26,7 +26,7 @@ Spawn advertises `{ outputSchema: true, depthLimit: true, toolFilter: true, pers #### What the model sees -The fresh child receives the standalone task content verbatim, inherits the parent model and workspace by default, and sees the global prompt with any configured child-scoped persona shadow. A tool filter removes global wire schemas, executable lookup, and Code Mode SDK bindings for that child but leaves independently registered guidance. It receives zero parent conversation messages; the filter is visibility/composition, not an authority grant inherited from the parent. +The fresh child receives the standalone task content verbatim, inherits the parent provider, model, reasoning effort, output-token limit, and workspace by default, and sees the global prompt with any configured child-scoped persona shadow. A tool filter removes global wire schemas, executable lookup, and Code Mode SDK bindings for that child but leaves independently registered guidance. It receives zero parent conversation messages; the filter is visibility/composition, not an authority grant inherited from the parent. #### Token effect @@ -52,4 +52,4 @@ Append-only; newly visible content follows the reusable request prefix and does ## Known Limitations and Deferred Work -- **Fresh means no parent transcript** — the child inherits cwd, lineage, model, and explicitly configured persona/tool restrictions, but none of the parent's conversation; use the fork provider when completed-turn context is required. +- **Fresh means no parent transcript** — the child inherits cwd, lineage, provider, model, reasoning effort, output-token limit, and explicitly configured persona/tool restrictions, but none of the parent's conversation; use the fork provider when completed-turn context is required. diff --git a/packages/subagent/subagent-spawn-in-process/README.zh.md b/packages/subagent/subagent-spawn-in-process/README.zh.md index 95a3b5cdb7..66ecbec2c0 100644 --- a/packages/subagent/subagent-spawn-in-process/README.zh.md +++ b/packages/subagent/subagent-spawn-in-process/README.zh.md @@ -6,13 +6,13 @@ spawn 提供方会在当前进程中创建一个全新的子 `Agent`。子 agent ## 行为 -`start(request)` 不传入 seed，直接委托给 [`startInProcessRun`](../subagent-in-process-driver/README.zh.md)，并在子 agent 发布后才返回。子 agent 获得父 agent 的工作目录/会话谱系，并默认继承父 agent 模型（除非覆盖），但以空对话开始运行。 +`start(request)` 不传入 seed，直接委托给 [`startInProcessRun`](../subagent-in-process-driver/README.zh.md)，并在子 agent 发布后才返回。子 agent 获得父 agent 的工作目录/会话谱系；除非 `request.agentOptions` 覆盖，否则还会继承父 agent 的提供方、模型、推理强度与输出 token 上限，但以空对话开始运行。 共享驱动器负责深度检查、persona 与工具过滤器设置、结构化输出、通过必需的信号执行取消、单次执行、结果读取和完全停稳后的 dispose（资源释放）。启动遭拒不会留下已发布的子 agent；启动调用兑现后卸载提供方，也不会撤销由持有方拥有的运行。 ## 能力 -spawn 声明 `{ outputSchema: true, depthLimit: true, toolFilter: true, persona: true }`，因为它控制子 agent 的创建窗口，能够强制执行全部四项功能。 +spawn 声明 `{ agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }`，因为它控制子 agent 的创建窗口，能够强制执行全部五项功能。 ## 配置 @@ -26,7 +26,7 @@ spawn 声明 `{ outputSchema: true, depthLimit: true, toolFilter: true, persona: #### 模型看到的内容 -全新的子 agent 逐字接收独立任务内容，默认继承父 agent 的模型和工作区，并看到带有已配置子 agent 作用域 persona 遮蔽的全局提示词。工具过滤器会为该子 agent 移除全局协议 schema、可执行工具查找和 Code Mode SDK 绑定，但保留独立注册的指导内容。它不接收任何父 agent 对话消息；过滤控制的是可见性与组合，并非从父 agent 继承的权限授予。 +全新的子 agent 逐字接收独立任务内容，默认继承父 agent 的提供方、模型、推理强度、输出 token 上限与工作区，并看到带有已配置子 agent 作用域 persona 遮蔽的全局提示词。工具过滤器会为该子 agent 移除全局协议 schema、可执行工具查找和 Code Mode SDK 绑定，但保留独立注册的指导内容。它不接收任何父 agent 对话消息；过滤控制的是可见性与组合，并非从父 agent 继承的权限授予。 #### Token 影响 @@ -52,4 +52,4 @@ spawn 声明 `{ outputSchema: true, depthLimit: true, toolFilter: true, persona: ## 已知限制与暂缓事项 -- **全新表示不含父 agent transcript（文本记录）**：子 agent 会继承 cwd、谱系、模型及显式配置的 persona/工具限制，但不继承父 agent 的任何对话；需要已完成轮次上下文时，请使用 fork 提供方。 +- **全新表示不含父 agent transcript（文本记录）**：子 agent 会继承 cwd、谱系、提供方、模型、推理强度、输出 token 上限及显式配置的 persona/工具限制，但不继承父 agent 的任何对话；需要已完成轮次上下文时，请使用 fork 提供方。 diff --git a/packages/subagent/subagent-spawn-in-process/src/index.ts b/packages/subagent/subagent-spawn-in-process/src/index.ts index dcd036e4ad..73811155c1 100644 --- a/packages/subagent/subagent-spawn-in-process/src/index.ts +++ b/packages/subagent/subagent-spawn-in-process/src/index.ts @@ -34,12 +34,18 @@ export const Config: z = z.object({ /** * The spawn provider. Supports every start-time capability: `depthLimit` (it * constructs the child, so it can enforce a recursion cap), `outputSchema` - * (the scoped structured runtime), and `toolFilter`/`persona` (scoped - * `restrict()` and a scoped shadowing persona section, applied in the child's - * creation window). + * (the scoped structured runtime), `agentOptions` (merged over the parent + * route), and `toolFilter`/`persona` (scoped `restrict()` and a scoped + * shadowing persona section, applied in the child's creation window). */ class SpawnInProcessProvider implements SubagentProvider { - readonly capabilities: SubagentCapabilities = { outputSchema: true, depthLimit: true, toolFilter: true, persona: true } + readonly capabilities: SubagentCapabilities = { + agentOptions: true, + outputSchema: true, + depthLimit: true, + toolFilter: true, + persona: true, + } // Context contract: a spawned child starts fresh — it never sees the parent conversation. readonly inheritsParentContext = false diff --git a/packages/subagent/subagent-spawn-in-process/tests/subagent-spawn-in-process.spec.ts b/packages/subagent/subagent-spawn-in-process/tests/subagent-spawn-in-process.spec.ts index ae60480c02..7a798a4818 100644 --- a/packages/subagent/subagent-spawn-in-process/tests/subagent-spawn-in-process.spec.ts +++ b/packages/subagent/subagent-spawn-in-process/tests/subagent-spawn-in-process.spec.ts @@ -282,10 +282,16 @@ describe('dsh-subagent-spawn-in-process', () => { await parentHandle.dispose() }) - it('advertises every start-time capability (depthLimit, outputSchema, toolFilter, persona)', async () => { + it('advertises every start-time capability', async () => { const { ctx } = await setup([]) const provider = ctx.subagents.getProvider('spawn')! - expect(provider.capabilities).toEqual({ outputSchema: true, depthLimit: true, toolFilter: true, persona: true }) + expect(provider.capabilities).toEqual({ + agentOptions: true, + outputSchema: true, + depthLimit: true, + toolFilter: true, + persona: true, + }) }) it('unregisters the provider when its fiber is disposed (HMR safety)', async () => { diff --git a/packages/subagent/subagent/README.i18n.yaml b/packages/subagent/subagent/README.i18n.yaml index a645351bd8..a55a61ee77 100644 --- a/packages/subagent/subagent/README.i18n.yaml +++ b/packages/subagent/subagent/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/subagent/README.md -README.md: e84a6b486253e81ccf7e7df12c4149e6df4ed9f2 -README.zh.md: e289863531c1686cedeccadfa76e2661dfa9bfc8 +README.md: 68ddc49197bcbd3f8eb5f362de60da33cb08c147 +README.zh.md: cf434152cd6366e371eef86f0edcb08d18978c66 diff --git a/packages/subagent/subagent/README.md b/packages/subagent/subagent/README.md index e84a6b4862..68ddc49197 100644 --- a/packages/subagent/subagent/README.md +++ b/packages/subagent/subagent/README.md @@ -26,7 +26,7 @@ The [subagent family overview](../README.md) maps implementations and model-faci | `listChildren(parentSessionId, signal?)` | List direct session-backed subagents with their `one-shot`/`continuable` mode, `running`/`inactive` activity, origin-classified one-level `hasChildren` hint, and per-child diagnostics, ordered by `createdAt` then id, without loading or resuming them. Reads the live session store and optional session persistence directly (live-only enumeration when persistence is absent) and requires the mounted `sessionProjections` registry; it does not require `ctx.agents`, the continuation manager, or any query service. | | `listDescendants(rootSessionId, signal?)` | Flatten the root's complete session tree in stable pre-order from the same live-preferred corpus, adding each subagent entry's durable `parentId` and root-relative `depth`. Ordinary sessions and one-shot children remain traversal nodes so continuable descendants below them are discovered. Identity, diagnostics, dependencies, and cancellation follow `listChildren()`. | -`SubagentStartRequest.label` is an optional short durable display label for a session-backed one-shot child. Model-facing delegation supplies its existing `description`; lower-level callers need not invent presentation metadata. Continuable starts always carry their own required label. `signal` is required and is the canonical cancellation channel for a one-shot `start`. An abort before publication makes `start()` reject after rollback; an abort after publication cancels the returned run's remaining turn work without hiding its id. The request may also select a model, require structured output, cap delegation depth, restrict child tools, or set a child persona. For a continuable start or follow-up, the caller signal owns lookup, materialization, and admission only until inbox acceptance; afterward the manager owns the Activation independently, so later caller cancellation neither cancels the accepted turn nor disposes the child. +`SubagentStartRequest.label` is an optional short durable display label for a session-backed one-shot child. Model-facing delegation supplies its existing `description`; lower-level callers need not invent presentation metadata. Continuable starts always carry their own required label. `signal` is required and is the canonical cancellation channel for a one-shot `start`. An abort before publication makes `start()` reject after rollback; an abort after publication cancels the returned run's remaining turn work without hiding its id. The request may also override the host Agent's provider, model, reasoning effort, and token limit, require structured output, cap delegation depth, restrict child tools, or set a child persona. Every requested optional feature requires its matching provider capability. For a continuable start or follow-up, the caller signal owns lookup, materialization, and admission only until inbox acceptance; afterward the manager owns the Activation independently, so later caller cancellation neither cancels the accepted turn nor disposes the child. Follow-up authority comes from the exact live direct parent recorded in the child's durable header. Cold resume checks that authority before reconstruction and again in the final no-await inbox-admission span, so a parent unregistered or replaced during materialization cannot authorize delivery. The `source` on a follow-up records who supplied the delivered message and grants no authority. @@ -36,11 +36,14 @@ Same-process requests, descriptors, results, and event payloads are trusted type Start-time features are advertised in `provider.capabilities` because the service must reject an unsupported one-shot request before child creation: +- `agentOptions` — apply host-Agent provider, model, reasoning-effort, and output-token overrides. - `outputSchema` — enforce a structured final result. - `depthLimit` — enforce `maxDepth`. - `toolFilter` — apply the requested child tool restriction. - `persona` — apply a per-child persona. +Both in-process providers advertise `agentOptions`: child creation merges requested fields over the provider, model, and reasoning effort in the parent's latest logged request, falling back to its creation options before the first request and retaining its configured token limit. A route change without an explicit effort clears the inherited route-owned effort so the selected model resolves its default. Current out-of-process providers advertise it as unsupported, so configured or model-selected overrides fail before their child transport starts instead of being silently ignored. + Every in-process child is composed by one call, `applyChildComposition(childCtx, parent, composition)`, which joins the parent's agent-preset composition before applying the child's own persona and tool filter. The join is what gives the child its capabilities: with every model-facing row on the agent plane, a child that joined nothing would reach the model with an empty tool registry ([`dsh-agent-presets`](../../preset/agent-presets/README.md)). Taking the parent as a parameter is deliberate — it makes composing a child WITHOUT that join unrepresentable at the call sites, which is the defect the one call exists to prevent. A deployment composing no preset roster joins nothing and needs nothing: its model-facing rows sit in the host composition, where the child already resolves them through the tool registry's global layer. `childSessionMeta()` records the joined preset id on the child's durable header for the same reason a top-level session records its own: the preset decides the tool schemas and prompt sections the model saw, so a cold read of the child's history has to rebuild that composition rather than the deployment default. It is read from the parent's live scope chain, not from the parent header, because a parent that switched preset while blank runs on the newer composition while its header still names the older one. @@ -49,7 +52,7 @@ Continuable creation is the optional `SubagentProvider.prepareContinuable?()` me ## The durable descriptor -The Service Definition owns the versioned `subagent/descriptor` session event vocabulary (`src/descriptor.ts`): `snapshotSubagentDescriptor()` validates and detaches the record before provider work, and `foldSubagentDescriptor()` validates the complete current-version payload before recovering it from a loaded child log. Every local session-backed start appends one descriptor with the provider name and lifecycle `mode`. A `one-shot` descriptor optionally carries the caller-owned durable display `label`; a `continuable` descriptor requires its durable creation label and additionally records resolved child `agentOptions.provider`/`model` and optional `persona`/`toolFilter` for cold resume. These are explicit fields, never the merge-extensible `AgentOptions` object, so an unrelated extension value cannot break continuation. The descriptor omits `subagentDepth` (the persisted header's `delegationDepth` is the monotone floor) and `outputSchema` (an Activation's result contract). The event is log-only: no `surfaceOp`, absent from model history, and retained by the append-only log across compaction. Malformed current-version payloads are corrupt; unsupported versions cannot be classified by this runtime. +The Service Definition owns the versioned `subagent/descriptor` session event vocabulary (`src/descriptor.ts`): `snapshotSubagentDescriptor()` validates and detaches the record before provider work, and `foldSubagentDescriptor()` validates the complete current-version payload before recovering it from a loaded child log. Every local session-backed start appends one descriptor with the provider name and lifecycle `mode`. A `one-shot` descriptor optionally carries the caller-owned durable display `label`; a `continuable` descriptor requires its durable creation label and additionally records resolved child `agentOptions.provider`/`model`/`reasoningEffort` and optional `persona`/`toolFilter` for cold resume. These are explicit fields, never the merge-extensible `AgentOptions` object, so an unrelated extension value cannot break continuation. The descriptor omits `subagentDepth` (the persisted header's `delegationDepth` is the monotone floor) and `outputSchema` (an Activation's result contract). The event is log-only: no `surfaceOp`, absent from model history, and retained by the append-only log across compaction. Malformed current-version payloads are corrupt; unsupported versions cannot be classified by this runtime. ## Delegation depth diff --git a/packages/subagent/subagent/README.zh.md b/packages/subagent/subagent/README.zh.md index e289863531..cf434152cd 100644 --- a/packages/subagent/subagent/README.zh.md +++ b/packages/subagent/subagent/README.zh.md @@ -26,7 +26,7 @@ subagent seam 允许一个 agent（智能体）通过具名提供方把工作委 | `listChildren(parentSessionId, signal?)` | 按 `createdAt`、再按 id 的顺序列出由会话支撑的直接 subagent，包括其 `one-shot`／`continuable` 模式、`running`／`inactive` 活动状态、根据 origin 分类得出的一层 `hasChildren` 提示，以及每个子级的诊断信息，且不会加载或恢复它们。该操作直接读取在线会话存储和可选的会话持久化（没有持久化时只枚举在线子级），并要求已挂载 `sessionProjections` 注册表；不要求 `ctx.agents`、继续执行管理器或任何查询服务。 | | `listDescendants(rootSessionId, signal?)` | 从同一份在线优先语料按稳定 pre-order 展平根的完整会话树，并为每个 subagent 条目附加持久 `parentId` 与相对根的 `depth`。普通会话与一次性 child 仍作为遍历节点，因此其下的可继续后代仍可发现。身份、diagnostic、依赖与取消约定均沿用 `listChildren()`。 | -`SubagentStartRequest.label` 是由会话支撑的一次性 child 所使用的可选简短持久化显示标签。面向模型的委派会提供其已有的 `description`；底层调用方无需凭空构造展示元数据。可继续启动始终携带自身的必填标签。`signal` 是必填项，也是一次性 `start` 的规范取消通道。发布前中止会使 `start()` 在回滚后拒绝；发布后中止会取消已返回 run 的剩余轮次工作，但不会隐藏其 id。请求还可以选择模型、要求结构化输出、限制委派深度、约束子 agent 工具或设置子 agent persona。对于可继续启动或后续操作，调用方信号只负责 inbox 接受前的查找、物化和准入；此后，Activation 由管理器独立拥有，因此调用方取消既不会取消已接受的轮次，也不会 dispose（资源释放）子 agent。 +`SubagentStartRequest.label` 是由会话支撑的一次性 child 所使用的可选简短持久化显示标签。面向模型的委派会提供其已有的 `description`；底层调用方无需凭空构造展示元数据。可继续启动始终携带自身的必填标签。`signal` 是必填项，也是一次性 `start` 的规范取消通道。发布前中止会使 `start()` 在回滚后拒绝；发布后中止会取消已返回 run 的剩余轮次工作，但不会隐藏其 id。请求还可以覆盖宿主 Agent 的提供方、模型、推理强度与 token 上限、要求结构化输出、限制委派深度、约束子 agent 工具或设置子 agent persona。每个被请求的可选特性都要求匹配的提供方能力。对于可继续启动或后续操作，调用方信号只负责 inbox 接受前的查找、物化和准入；此后，Activation 由管理器独立拥有，因此调用方取消既不会取消已接受的轮次，也不会 dispose（资源释放）子 agent。 后续操作的权限来自子 agent 持久化 header 中记录的确切在线直接父级。冷恢复会在重建前检查该权限，并在最终无 await 的 inbox 准入区间再次检查，因此在物化期间被注销或替换的 parent 无法授权投递。后续操作上的 `source` 记录谁提供了所投递的消息，不授予任何权限。 @@ -36,11 +36,14 @@ subagent seam 允许一个 agent（智能体）通过具名提供方把工作委 启动时功能通过 `provider.capabilities` 声明，因为服务必须在创建子 agent 前拒绝不受支持的一次性请求： +- `agentOptions`：应用宿主 Agent 提供方、模型、推理强度与输出 token 上限覆盖； - `outputSchema`：强制执行结构化最终结果； - `depthLimit`：强制执行 `maxDepth`； - `toolFilter`：应用请求的子 agent 工具限制； - `persona`：应用每个子 agent 独立的 persona。 +两个进程内提供方都会声明 `agentOptions`：创建子 agent 时，请求字段会覆盖父级最新记录请求中的提供方、模型与推理强度；首个请求之前回退到其创建选项，并保留其中配置的 token 上限。更换路由但没有显式指定强度时，会清除继承的路由所属强度，使所选模型解析自己的默认值。当前进程外提供方会声明不支持，因此配置或模型选择的覆盖会在启动子传输前失败，而不会被静默忽略。 + 每个进程内子 agent 都通过一次 `applyChildComposition(childCtx, parent, composition)` 调用完成组装：先加入父级的 agent-preset 组合，再应用子 agent 自己的 persona 和工具限制。加入父级组合正是子 agent 获得能力的途径：所有面向模型的行都位于 agent 平面，完全没有加入任何组合的子 agent 抵达模型时会看到空的工具注册表（见 [`dsh-agent-presets`](../../preset/agent-presets/README.zh.md)）。将父级作为参数是刻意设计：这让“组装子 agent 却不做该加入”在各调用点无法表达，而这正是这一次调用所要杜绝的缺陷。未组装 preset roster 的部署不加入任何组合、也不需要加入；其面向模型的行位于宿主组合中，子 agent 已能通过工具注册表的全局层解析到它们。 `childSessionMeta()` 把所加入的 preset id 记在子 agent 的持久化 header 上，理由与顶层会话记录自己的那一个相同：preset 决定了模型所见的工具 schema 与提示段，因此冷读子 agent 的历史时必须重建那份组装，而不是部署默认值。该值从父方**活着的** scope 链读取，而不是从父方 header 读取，因为在空白期切换过 preset 的父方运行在更新的那份组装上，而它的 header 仍写着旧的那个。 @@ -49,7 +52,7 @@ subagent seam 允许一个 agent（智能体）通过具名提供方把工作委 ## 持久化描述符 -该 Service Definition 拥有版本化的 `subagent/descriptor` 会话事件词汇（`src/descriptor.ts`）：`snapshotSubagentDescriptor()` 会在提供方工作之前校验并分离记录，`foldSubagentDescriptor()` 则会在从已加载子 agent 日志中恢复描述符之前，校验当前版本的完整 payload。每次由本地会话支撑的启动都会追加一个带有提供方名称与生命周期 `mode` 的描述符。`one-shot` 描述符可以携带调用方拥有的可选持久化显示 `label`；`continuable` 描述符要求其持久化创建标签，并另外记录已解析的子 agent `agentOptions.provider`／`model`，以及用于从持久化存储恢复的可选 `persona`／`toolFilter`。这些是显式字段，绝不是可通过合并扩展的 `AgentOptions` 对象，因此无关的扩展值不会破坏继续执行。描述符省略 `subagentDepth`（持久化 header 的 `delegationDepth` 是单调下界）和 `outputSchema`（单次 Activation 的结果约定）。该事件只进入日志：不含 `surfaceOp`，不进入模型历史，并由仅追加日志跨压缩（compaction）保留。格式错误的当前版本 payload 属于损坏；本运行时无法对不受支持的版本进行分类。 +该 Service Definition 拥有版本化的 `subagent/descriptor` 会话事件词汇（`src/descriptor.ts`）：`snapshotSubagentDescriptor()` 会在提供方工作之前校验并分离记录，`foldSubagentDescriptor()` 则会在从已加载子 agent 日志中恢复描述符之前，校验当前版本的完整 payload。每次由本地会话支撑的启动都会追加一个带有提供方名称与生命周期 `mode` 的描述符。`one-shot` 描述符可以携带调用方拥有的可选持久化显示 `label`；`continuable` 描述符要求其持久化创建标签，并另外记录已解析的子 agent `agentOptions.provider`／`model`／`reasoningEffort`，以及用于从持久化存储恢复的可选 `persona`／`toolFilter`。这些是显式字段，绝不是可通过合并扩展的 `AgentOptions` 对象，因此无关的扩展值不会破坏继续执行。描述符省略 `subagentDepth`（持久化 header 的 `delegationDepth` 是单调下界）和 `outputSchema`（单次 Activation 的结果约定）。该事件只进入日志：不含 `surfaceOp`，不进入模型历史，并由仅追加日志跨压缩（compaction）保留。格式错误的当前版本 payload 属于损坏；本运行时无法对不受支持的版本进行分类。 ## 委派深度 diff --git a/packages/subagent/subagent/src/child-agent.ts b/packages/subagent/subagent/src/child-agent.ts index 7582338858..22c9e77bf5 100644 --- a/packages/subagent/subagent/src/child-agent.ts +++ b/packages/subagent/subagent/src/child-agent.ts @@ -57,9 +57,38 @@ export function resolveChildDepth(parent: Agent, maxDepth: number | undefined): } /** - * Resolve the child's `AgentOptions`: the parent's provider/model/maxTokens - * route unless the request overrides it, stamped with the child's own - * delegation depth. + * Resolve the parent values inherited by a child. The latest request header + * owns provider, model, and reasoning effort after request-time selection; + * creation options remain the fallback before the first request and retain + * the configured output-token limit. + * @param parent - delegating parent Agent. + * @returns detached Agent options for child-option merging. + */ +export function parentAgentOptionsForDelegation(parent: Agent): AgentOptions { + const requestConfig = parent.session.requestHeader()?.config + if (requestConfig === undefined) return { ...parent.options } + const { + provider: _createdProvider, + model: _createdModel, + reasoningEffort: _createdReasoningEffort, + ...createdOptions + } = parent.options + return { + ...createdOptions, + provider: requestConfig.provider, + model: requestConfig.model, + ...requestConfig.reasoningEffort === undefined + ? {} + : { reasoningEffort: requestConfig.reasoningEffort }, + } +} + +/** + * Resolve the child's `AgentOptions`: the parent's provider/model, + * reasoning-effort, and maxTokens values unless the request overrides them, + * stamped with the child's own delegation depth. Changing the route without + * naming an effort clears the parent's route-owned effort so the selected + * model resolves its own default. * @param parent - the delegating parent whose route the child inherits. * @param requested - per-child overrides, if any. * @param childDepth - the resolved delegation depth to stamp. @@ -70,16 +99,22 @@ export function resolveChildAgentOptions( requested: AgentOptions | undefined, childDepth: number, ): AgentOptions { - const parentProvider = parent.options.provider - const parentModel = parent.options.model - const parentMaxTokens = parent.options.maxTokens - return { + const parentOptions = parentAgentOptionsForDelegation(parent) + const parentProvider = parentOptions.provider + const parentModel = parentOptions.model + const parentReasoningEffort = parentOptions.reasoningEffort + const parentMaxTokens = parentOptions.maxTokens + const resolved: AgentOptions = { ...parentProvider !== undefined ? { provider: parentProvider } : {}, ...parentModel !== undefined ? { model: parentModel } : {}, + ...parentReasoningEffort !== undefined ? { reasoningEffort: parentReasoningEffort } : {}, ...parentMaxTokens !== undefined ? { maxTokens: parentMaxTokens } : {}, ...requested, subagentDepth: childDepth, } + const routeChanged = resolved.provider !== parentProvider || resolved.model !== parentModel + if (routeChanged && requested?.reasoningEffort === undefined) delete resolved.reasoningEffort + return resolved } /** diff --git a/packages/subagent/subagent/src/continuation.ts b/packages/subagent/subagent/src/continuation.ts index 652a3ba6c8..2588c1a699 100644 --- a/packages/subagent/subagent/src/continuation.ts +++ b/packages/subagent/subagent/src/continuation.ts @@ -30,7 +30,7 @@ import type { AgentSetupCommit, CreateAgentOptions, } from '@deepseek-ai/dsh-agent' -import { boundContextSummary, createUserMessage, errorChain } from '@deepseek-ai/dsh-llm' +import { ReasoningEffortId, boundContextSummary, createUserMessage, errorChain } from '@deepseek-ai/dsh-llm' import type { ContentBlock, MessageId, MessageSource } from '@deepseek-ai/dsh-llm' import { SessionId } from '@deepseek-ai/dsh-session' import type { SessionEvent } from '@deepseek-ai/dsh-session' @@ -417,14 +417,17 @@ export class SubagentContinuationManager { const childDepth = resolveChildDepth(parent, request.maxDepth) // Snapshot before any await: invalid descriptor JSON rejects the call // before a child exists, and the detached value is what reaches the log. - const agentProvider = request.agentOptions?.provider ?? parent.options.provider - const agentModel = request.agentOptions?.model ?? parent.options.model + const agentOptions = resolveChildAgentOptions(parent, request.agentOptions, childDepth) + const agentProvider = agentOptions.provider + const agentModel = agentOptions.model + const agentReasoningEffort = agentOptions.reasoningEffort const descriptor = snapshotSubagentDescriptor({ mode: 'continuable', provider: spec.provider, label: spec.label, ...agentProvider !== undefined ? { agentProvider } : {}, ...agentModel !== undefined ? { agentModel } : {}, + ...agentReasoningEffort !== undefined ? { agentReasoningEffort } : {}, ...request.persona !== undefined ? { persona: request.persona } : {}, ...request.toolFilter !== undefined ? { toolFilter: request.toolFilter } : {}, }) @@ -460,7 +463,7 @@ export class SubagentContinuationManager { provider: spec.provider, parent, create: { seed, meta: childSessionMeta(parent, childDepth, lineageSeedLength), delegatedPolicies }, - agentOptions: resolveChildAgentOptions(parent, request.agentOptions, childDepth), + agentOptions, composition: { persona: request.persona, toolFilter: request.toolFilter }, signal: spec.signal, }) @@ -981,6 +984,9 @@ export class SubagentContinuationManager { agentOptions: { ...descriptor.agentProvider !== undefined ? { provider: descriptor.agentProvider } : {}, ...descriptor.agentModel !== undefined ? { model: descriptor.agentModel } : {}, + ...descriptor.agentReasoningEffort !== undefined + ? { reasoningEffort: ReasoningEffortId(descriptor.agentReasoningEffort) } + : {}, }, composition: { persona: descriptor.persona, toolFilter: descriptor.toolFilter }, signal: options.signal, diff --git a/packages/subagent/subagent/src/descriptor.ts b/packages/subagent/subagent/src/descriptor.ts index 6d9dedee75..9a25c382b1 100644 --- a/packages/subagent/subagent/src/descriptor.ts +++ b/packages/subagent/subagent/src/descriptor.ts @@ -23,6 +23,7 @@ import { snapshotJsonValue } from '@deepseek-ai/dsh-session' import type { SessionEvent } from '@deepseek-ai/dsh-session' +import type { ReasoningEffortId } from '@deepseek-ai/dsh-llm' import type { ToolRestriction } from '@deepseek-ai/dsh-tools' declare module '@deepseek-ai/dsh-session/types' { @@ -44,7 +45,7 @@ declare module '@deepseek-ai/dsh-session/types' { * Supporting another composition input is a deliberate version change, never * an implicit extra field. */ -export const SUBAGENT_DESCRIPTOR_VERSION = 2 +export const SUBAGENT_DESCRIPTOR_VERSION = 3 /** Fields shared by every supported `subagent/descriptor` payload. */ interface SubagentDescriptorBase { @@ -76,6 +77,8 @@ export interface ContinuableSubagentDescriptorData extends SubagentDescriptorBas readonly agentProvider?: string /** Resolved child `agentOptions.model`, when one was declared. */ readonly agentModel?: string + /** Resolved child `agentOptions.reasoningEffort`, when one was declared. */ + readonly agentReasoningEffort?: ReasoningEffortId /** Per-child persona that shadows the deployment persona on resume. */ readonly persona?: string /** Child tool scoping reapplied on resume. */ @@ -111,6 +114,8 @@ export interface ContinuableSubagentDescriptorInput extends SubagentDescriptorIn readonly agentProvider?: string /** Requested child `agentOptions.model`. */ readonly agentModel?: string + /** Requested child `agentOptions.reasoningEffort`. */ + readonly agentReasoningEffort?: ReasoningEffortId /** Requested per-child persona. */ readonly persona?: string /** Requested child tool scoping. */ @@ -133,6 +138,7 @@ const CONTINUABLE_DESCRIPTOR_KEYS = new Set([ ...DESCRIPTOR_BASE_KEYS, 'agentProvider', 'agentModel', + 'agentReasoningEffort', 'persona', 'toolFilter', ]) @@ -231,6 +237,7 @@ function parseSubagentDescriptor(value: unknown): SubagentDescriptorData | undef } const agentProvider = optionalString(value, 'agentProvider') const agentModel = optionalString(value, 'agentModel') + const agentReasoningEffort = optionalString(value, 'agentReasoningEffort') as ReasoningEffortId | undefined const persona = optionalString(value, 'persona') const toolFilter = Object.hasOwn(value, 'toolFilter') ? parseToolFilter(value['toolFilter']) @@ -242,6 +249,7 @@ function parseSubagentDescriptor(value: unknown): SubagentDescriptorData | undef label, ...agentProvider !== undefined ? { agentProvider } : {}, ...agentModel !== undefined ? { agentModel } : {}, + ...agentReasoningEffort !== undefined ? { agentReasoningEffort } : {}, ...persona !== undefined ? { persona } : {}, ...toolFilter !== undefined ? { toolFilter } : {}, } @@ -283,6 +291,7 @@ export function snapshotSubagentDescriptor(input: SubagentDescriptorInput): Suba label: input.label, ...input.agentProvider !== undefined ? { agentProvider: input.agentProvider } : {}, ...input.agentModel !== undefined ? { agentModel: input.agentModel } : {}, + ...input.agentReasoningEffort !== undefined ? { agentReasoningEffort: input.agentReasoningEffort } : {}, ...input.persona !== undefined ? { persona: input.persona } : {}, ...input.toolFilter !== undefined ? { toolFilter: input.toolFilter } : {}, } diff --git a/packages/subagent/subagent/src/index.ts b/packages/subagent/subagent/src/index.ts index 2f29e32010..42dca84908 100644 --- a/packages/subagent/subagent/src/index.ts +++ b/packages/subagent/subagent/src/index.ts @@ -103,6 +103,7 @@ export { applyChildComposition, captureDelegatedPolicyOverrides, childSessionMeta, + parentAgentOptionsForDelegation, resolveChildAgentOptions, resolveChildDepth, SubagentDepthError, @@ -494,6 +495,7 @@ export class SubagentRuntime extends Service { /** Reject the first requested capability that the provider lacks. */ private assertCapabilities(provider: SubagentProvider, request: SubagentStartRequest): void { const needs: { when: boolean; cap: keyof SubagentCapabilities }[] = [ + { when: request.agentOptions !== undefined, cap: 'agentOptions' }, { when: request.outputSchema !== undefined, cap: 'outputSchema' }, { when: request.maxDepth !== undefined, cap: 'depthLimit' }, { when: request.toolFilter !== undefined, cap: 'toolFilter' }, diff --git a/packages/subagent/subagent/src/out-of-process.ts b/packages/subagent/subagent/src/out-of-process.ts index abb6dd50e7..2667884af4 100644 --- a/packages/subagent/subagent/src/out-of-process.ts +++ b/packages/subagent/subagent/src/out-of-process.ts @@ -44,10 +44,11 @@ function limitSubagentDiagnostic(diagnostic: string): string { /** * The capability advertisement of an out-of-process backend: NONE. A child in * another process cannot honor parent-enforced start features - * (`outputSchema`/`maxDepth`/`toolFilter`/`persona`), so the service rejects a + * (`agentOptions`/`outputSchema`/`maxDepth`/`toolFilter`/`persona`), so the service rejects a * request needing any of them before `start` runs — never accepted-then-ignored. */ export const NO_START_CAPABILITIES: SubagentCapabilities = Object.freeze({ + agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, diff --git a/packages/subagent/subagent/src/types.ts b/packages/subagent/subagent/src/types.ts index 17978550ab..415379ca75 100644 --- a/packages/subagent/subagent/src/types.ts +++ b/packages/subagent/subagent/src/types.ts @@ -84,6 +84,7 @@ export interface SubagentRunEndInfo { * to `maxDepth`; the other names match. */ export interface SubagentCapabilities { + readonly agentOptions: boolean readonly outputSchema: boolean readonly depthLimit: boolean readonly toolFilter: boolean @@ -116,6 +117,12 @@ export interface SubagentStartRequest { * remaining turn work when it fires afterward. */ readonly signal: AbortSignal + /** + * Optional host-Agent provider, model, reasoning-effort, and output-token + * overrides. Requires {@link SubagentCapabilities.agentOptions}; in-process + * providers merge them over the parent Agent's options when they create the + * child. + */ readonly agentOptions?: AgentOptions /** * Object-rooted JSON Schema within `assertObjectJsonSchema`'s enforced subset. Start rejects diff --git a/packages/subagent/subagent/tests/child-agent.spec.ts b/packages/subagent/subagent/tests/child-agent.spec.ts new file mode 100644 index 0000000000..92302cfca4 --- /dev/null +++ b/packages/subagent/subagent/tests/child-agent.spec.ts @@ -0,0 +1,76 @@ +import { describe, expect, it } from 'vitest' +import type { Agent } from '@deepseek-ai/dsh-agent' +import { ReasoningEffortId } from '@deepseek-ai/dsh-llm' +import { Session, SessionId } from '@deepseek-ai/dsh-session' +import { resolveChildAgentOptions } from '../src/child-agent.ts' + +function parentAgent(): Agent { + const id = SessionId('parent') + return { + id, + options: { + provider: 'parent-provider', + model: 'parent-model', + reasoningEffort: ReasoningEffortId('high'), + maxTokens: 512, + }, + session: Session.create(id), + } as Agent +} + +describe('child Agent options', () => { + it('inherits the parent effort while the exact route is unchanged', () => { + expect(resolveChildAgentOptions(parentAgent(), undefined, 1)).toEqual({ + provider: 'parent-provider', + model: 'parent-model', + reasoningEffort: 'high', + maxTokens: 512, + subagentDepth: 1, + }) + }) + + it('clears an inherited effort when the child route changes', () => { + expect(resolveChildAgentOptions(parentAgent(), { model: 'child-model' }, 1)).toEqual({ + provider: 'parent-provider', + model: 'child-model', + maxTokens: 512, + subagentDepth: 1, + }) + }) + + it('keeps an explicit child effort when the child route changes', () => { + expect(resolveChildAgentOptions(parentAgent(), { + provider: 'child-provider', + model: 'child-model', + reasoningEffort: ReasoningEffortId('max'), + }, 1)).toEqual({ + provider: 'child-provider', + model: 'child-model', + reasoningEffort: 'max', + maxTokens: 512, + subagentDepth: 1, + }) + }) + + it('inherits the latest logged request selection over creation-time values', () => { + const parent = parentAgent() + parent.session.append('request/header', { + header: { + config: { + provider: 'current-provider', + model: 'current-model', + reasoningEffort: ReasoningEffortId('low'), + }, + }, + reason: 'initial', + }) + + expect(resolveChildAgentOptions(parent, undefined, 1)).toEqual({ + provider: 'current-provider', + model: 'current-model', + reasoningEffort: 'low', + maxTokens: 512, + subagentDepth: 1, + }) + }) +}) diff --git a/packages/subagent/subagent/tests/continuation.spec.ts b/packages/subagent/subagent/tests/continuation.spec.ts index 6cf1aea5a8..d1d40e0ff4 100644 --- a/packages/subagent/subagent/tests/continuation.spec.ts +++ b/packages/subagent/subagent/tests/continuation.spec.ts @@ -12,7 +12,7 @@ import JsonlSessionPersistence from '@deepseek-ai/dsh-session-persistence-jsonl' import * as SubagentSpawn from '@deepseek-ai/dsh-subagent-spawn-in-process' import * as SubagentFork from '@deepseek-ai/dsh-subagent-fork-in-process' import type { GenerateOptions, MessageId, StreamChunk } from '@deepseek-ai/dsh-llm' -import { CallId, createUserMessage, LlmAdapter } from '@deepseek-ai/dsh-llm' +import { CallId, createUserMessage, LlmAdapter, ReasoningEffortId } from '@deepseek-ai/dsh-llm' import { defineTool } from '@deepseek-ai/dsh-tools' import InvariantRegistry from '@deepseek-ai/dsh-invariants' import { MockAdapter, maxTokensResponse, textResponse, toolCallResponse } from '../../../core/agent-loop/tests/mock-adapter.ts' @@ -239,7 +239,7 @@ describe('SubagentRuntime.startContinuable', () => { const start = vi.fn(async () => { throw new Error('must not dispatch') }) ctx.subagents.registerProvider({ name: 'one-shot', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start, }) @@ -283,6 +283,40 @@ describe('SubagentRuntime.startContinuable', () => { expect(loaded.meta.origin).toBe('subagent') }) + it('persists a selected reasoning effort and reapplies it on cold resume', async () => { + const effort = ReasoningEffortId('max') + const adapter = new MockAdapter([ + textResponse('first answer'), + textResponse('resumed answer'), + ], { + efforts: [{ id: effort, name: 'Max' }], + defaultEffort: effort, + }) + const { ctx, parent } = await setupWith(adapter) + parkParent(ctx, parent) + const childEfforts: Array = [] + ctx.on('agent/created', ({ agent }) => { + if (agent !== parent) childEfforts.push(agent.options.reasoningEffort) + }) + + const started = await ctx.subagents.startContinuable({ + ...startSpec(parent), + request: { + prompt: message('selected reasoning'), + parent, + agentOptions: { reasoningEffort: effort }, + }, + }) + await waitNoActivation(ctx, started.childId) + const loaded = await ctx.sessionPersistence.load(started.childId) + expect(loaded.events.find(event => event.type === 'subagent/descriptor')?.data) + .toMatchObject({ agentReasoningEffort: 'max' }) + + await followup(ctx, parent, started.childId, message('resume selected reasoning')) + await waitNoActivation(ctx, started.childId) + expect(childEfforts).toEqual(['max', 'max']) + }) + it('rolls the child back completely when the caller signal aborts before acceptance', async () => { const { ctx, parent } = await setup([textResponse('unused')]) const controller = new AbortController() @@ -531,7 +565,7 @@ describe('SubagentRuntime.followup residency routing', () => { await ctx.plugin(SubagentInvariant) const disposeProvider = ctx.subagents.registerProvider({ name: 'retired', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => { throw new Error('one-shot start is not used') }, prepareContinuable: () => Promise.resolve({}), @@ -2418,27 +2452,43 @@ describe('continuable errors', () => { hold.resolve(undefined) }) - it('reapplies the descriptor model route on cold resume', async () => { - const { ctx, parent } = await setup([textResponse('first'), textResponse('resumed')]) + it('reapplies the descriptor model route and reasoning effort on cold resume', async () => { + const effort = ReasoningEffortId('high') + const adapter = new MockAdapter([textResponse('first'), textResponse('resumed')], { + efforts: [{ id: effort, name: 'High' }], + defaultEffort: effort, + }) + const { ctx, parent } = await setupWith(adapter) const started = await ctx.subagents.startContinuable({ ...startSpec(parent), request: { prompt: message('routed work'), parent, - agentOptions: { provider: 'mock', model: 'child-model' }, + agentOptions: { provider: 'mock', model: 'child-model', reasoningEffort: effort }, }, }) await waitNoActivation(ctx, started.childId) const loaded = await ctx.sessionPersistence.load(started.childId) expect(loaded.events.find(event => event.type === 'subagent/descriptor')?.data) - .toMatchObject({ agentProvider: 'mock', agentModel: 'child-model' }) + .toMatchObject({ + agentProvider: 'mock', + agentModel: 'child-model', + agentReasoningEffort: 'high', + }) // The resumed Activation runs on the declared route, not the parent's. await followup(ctx, parent, started.childId, message('again')) await vi.waitFor(() => { - expect(ctx.agents.get(started.childId)?.options.model).toBe('child-model') + expect(ctx.agents.get(started.childId)?.options).toMatchObject({ + model: 'child-model', + reasoningEffort: 'high', + }) }) await waitNoActivation(ctx, started.childId) + const resumed = await ctx.sessionPersistence.load(started.childId) + expect(resumed.events.flatMap(event => event.type === 'request/header' + ? [event.data.header.config.reasoningEffort] + : [])).toEqual([effort, effort]) }) it('unloading the manager drains its live activations', async () => { diff --git a/packages/subagent/subagent/tests/invariant.spec.ts b/packages/subagent/subagent/tests/invariant.spec.ts index 91200abf8d..08792c2348 100644 --- a/packages/subagent/subagent/tests/invariant.spec.ts +++ b/packages/subagent/subagent/tests/invariant.spec.ts @@ -21,7 +21,7 @@ async function setup(): Promise { const provider = (name: string): SubagentProvider => ({ name, - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => { throw new Error('not used') }, }) diff --git a/packages/subagent/subagent/tests/out-of-process.spec.ts b/packages/subagent/subagent/tests/out-of-process.spec.ts index 0d3307ca51..98b670f36f 100644 --- a/packages/subagent/subagent/tests/out-of-process.spec.ts +++ b/packages/subagent/subagent/tests/out-of-process.spec.ts @@ -21,7 +21,13 @@ import { describe('NO_START_CAPABILITIES', () => { it('advertises nothing and is frozen (shared by every out-of-process backend)', () => { - expect(NO_START_CAPABILITIES).toEqual({ outputSchema: false, depthLimit: false, toolFilter: false, persona: false }) + expect(NO_START_CAPABILITIES).toEqual({ + agentOptions: false, + outputSchema: false, + depthLimit: false, + toolFilter: false, + persona: false, + }) expect(Object.isFrozen(NO_START_CAPABILITIES)).toBe(true) }) }) diff --git a/packages/subagent/subagent/tests/service.spec.ts b/packages/subagent/subagent/tests/service.spec.ts index 05e9785611..6a93f2fdef 100644 --- a/packages/subagent/subagent/tests/service.spec.ts +++ b/packages/subagent/subagent/tests/service.spec.ts @@ -2,7 +2,7 @@ import { describe, expect, expectTypeOf, it, vi } from 'vitest' import { Context } from '@deepseek-ai/cordis' import { type Agent } from '@deepseek-ai/dsh-agent' -import { HarnessError } from '@deepseek-ai/dsh-llm' +import { HarnessError, ReasoningEffortId } from '@deepseek-ai/dsh-llm' import { carrierKeyOf } from '@deepseek-ai/dsh-scope' import SubagentRuntime, { foldSubagentDescriptor, @@ -24,8 +24,8 @@ function fakeParent(id = 'parent-1'): Agent { return { id: SessionId(id) } as unknown as Agent } -const ALL_CAPS: SubagentCapabilities = { outputSchema: true, depthLimit: true, toolFilter: true, persona: true } -const NO_CAPS: SubagentCapabilities = { outputSchema: false, depthLimit: false, toolFilter: false, persona: false } +const ALL_CAPS: SubagentCapabilities = { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: true } +const NO_CAPS: SubagentCapabilities = { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false } function baseRequest(overrides: Partial = {}): SubagentStartRequest { return { @@ -162,6 +162,7 @@ describe('SubagentRuntime', () => { }) it.each([ + ['agentOptions', { agentOptions: { model: 'child-model' } }], ['outputSchema', { outputSchema: { type: 'object', properties: {} } }], ['depthLimit', { maxDepth: 1 }], ['toolFilter', { toolFilter: { deny: ['bash'] } }], @@ -348,6 +349,7 @@ describe('subagent descriptors', () => { label: 'complete child', agentProvider: 'deepseek', agentModel: 'chat', + agentReasoningEffort: ReasoningEffortId('high'), persona: 'reviewer', toolFilter: { allow: ['read'], deny: ['bash'] }, } @@ -357,6 +359,7 @@ describe('subagent descriptors', () => { label: complete.label, agentProvider: complete.agentProvider, agentModel: complete.agentModel, + agentReasoningEffort: complete.agentReasoningEffort, persona: complete.persona, toolFilter: complete.toolFilter, })).toEqual(complete) @@ -448,6 +451,13 @@ describe('subagent descriptors', () => { label: 'l', agentModel: [], }, 'agentModel must be a string'], + ['invalid agent reasoning effort', { + version: SUBAGENT_DESCRIPTOR_VERSION, + mode: 'continuable', + provider: 'spawn', + label: 'l', + agentReasoningEffort: 7, + }, 'agentReasoningEffort must be a string'], ['invalid persona', { version: SUBAGENT_DESCRIPTOR_VERSION, mode: 'continuable', diff --git a/packages/subagent/tool-subagent/README.i18n.yaml b/packages/subagent/tool-subagent/README.i18n.yaml index 29d65cf9da..f6caeb1941 100644 --- a/packages/subagent/tool-subagent/README.i18n.yaml +++ b/packages/subagent/tool-subagent/README.i18n.yaml @@ -2,5 +2,5 @@ # side as of the last confirmed-consistent state. Both languages carry equal authority; # after editing either side, bring the other along and re-record with: # pnpm run verify-translation-pairing --write packages/subagent/tool-subagent/README.md -README.md: 28e6213b903ffffa7934e244b2a74ada519b32b2 -README.zh.md: 74e8896a152c787abd0aebf055d6e13f6158bcd3 +README.md: e643442de7fa45f15a5c2bf818e2c25feb44b6c5 +README.zh.md: aa6dec73c66ce6b4db525d09cd166e671dbec9dc diff --git a/packages/subagent/tool-subagent/README.md b/packages/subagent/tool-subagent/README.md index 28e6213b90..e643442de7 100644 --- a/packages/subagent/tool-subagent/README.md +++ b/packages/subagent/tool-subagent/README.md @@ -6,7 +6,11 @@ The model-facing delegation tool over one configured `ctx.subagents` provider. C ## Provider selection and lifecycle -Each plugin instance binds one `provider` to one `toolName`; the model receives no provider selector. Load another distinctly named instance to expose another transport. The tool registers only while its provider exists, avoiding sibling load-order and provider-reload dependencies. Its description follows `provider.inheritsParentContext`: fresh children require standalone prompts, while forked children already see completed parent turns. +Each plugin instance binds one subagent transport `provider` to one `toolName`; the model cannot change that transport. Load another distinctly named instance to expose another transport. `enableModelSelection: true`, or an enabled Host preference when `modelSelectionSettings: true`, requires that provider's child `agentOptions` capability and exposes optional child LLM `provider`, `model`, and `reasoning_effort` fields without additional route configuration. A call may supply a complete provider/model pair, or only an effort when configured or parent values supply the effective route. The live adapter resolves explicit or configured routes before child creation. A call that omits every selection field uses `agentOptions` and then inherits compatible missing values from the parent's latest logged request selection, falling back to its creation options before the first request and retaining its configured `maxTokens`. Changing provider or model without naming an effort clears the lower layer's route-owned effort so the selected model resolves its default. + +The delegation tool registers only while its subagent provider exists, avoiding sibling load-order and provider-reload dependencies. When model selection is enabled, its optional fields remain visible without `ctx.llm`; a call that selects a route rejects if the service is unavailable. When disabled, the schema omits those fields and execution rejects a forced selection. Configured `agentOptions` remain deployment-owned child defaults independently of this model-facing switch. Adapter catalog and topology changes do not rewrite or re-register the tool. Its description follows `provider.inheritsParentContext`: fresh children require standalone prompts, while forked children already see completed parent turns. + +An enabled definition registers `list_subagent_models`, which lists registered providers, one provider's advertised models, or one exact model's reasoning efforts at call time. At most one instance in a tool scope may enable selection because this discovery tool has a global name; duplicate owners fail registration. Shipped product compositions default the primary `subagent` (`spawn`) instance off and sample the Host `subagent-model-selection.enabled` preference when each new top-level session is composed. The enabled decision is logged as `subagent/model-selection-enabled`, inherited by child sessions, and retained on resume; later settings edits do not change a running session. Shipped compositions deliberately keep `subagent_fork` disabled so the fork inherits the parent's provider and model: changing that route would forfeit provider-side KV Cache reuse of the inherited conversation prefix and can make prefix recomputation dominate the delegated task's cost. This restriction remains even if discovery ownership is separated. Catalog membership remains advisory: an enabled delegation tool accepts an unlisted model id when its adapter does. The [model-selected route Agent Note](../../../.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.md) owns the rationale and reintroduction condition. A foreground call passes the execution signal through startup and execution, awaits `run.result`, and always awaits `run.dispose()` before returning. Only `completed` returns the canonical `{ kind: 'foreground', runId, output: JsonValue[] }`, rendered as the same final text. Abort, refusal, token limit, and other failures become errored tool results whose message contains the stop-reason headline, an optional provider-authored `SubagentResult.diagnostic`, and then any preserved partial assistant text. The diagnostic remains separate from `SubagentResult.output`, so a truncated answer is never reported as success or confused with infrastructure detail. If result collection and disposal both reject, the errored result preserves both failures. @@ -20,9 +24,11 @@ A foreground call passes the execution signal through startup and execution, awa |---|---| | `provider` (required) | Provider name (`spawn`, `fork`, `acp`, ...). | | `toolName` | Model-facing name, default `subagent`; distinct for every loaded instance. | +| `enableModelSelection` | Exposes and accepts model-facing child LLM selection fields and registers the shared `list_subagent_models` tool, default `false`. It requires the subagent provider's `agentOptions` capability. At most one instance in a tool scope may enable it; the discovery schema remains registered without `ctx.llm`, while discovery and selected-route calls reject until that optional service is available. Configured `agentOptions` remain available when this switch is disabled. | +| `modelSelectionSettings` | Samples the Host `subagent-model-selection` preference while composing an Agent, records an enabled decision in its Session, and inherits that decision in child Sessions. Default `false`; mutually exclusive with `enableModelSelection` and valid only in an Agent-scoped composition. The preference defaults off and changes only subsequently composed top-level Sessions. | | `enableRunInBackground` | Exposes background mode, default `true`; disabling also rejects forced background calls. | | `backgroundMode` | Background lifecycle policy, default `one-shot`. `one-shot` defaults calls to foreground; `continuable` defaults them to background, requires the provider's `prepareContinuable` capability, and returns a durable child id without requiring the follow-up tool. | -| `agentOptions` | Provider-specific child `provider`, `model`, and positive `maxTokens`; the in-process provider treats explicit values as overrides of inherited parent options. | +| `agentOptions` | Configured child LLM `provider`, `model`, adapter-owned `reasoningEffort`, and positive `maxTokens`; requires the subagent provider's `agentOptions` capability. In-process providers merge explicit values over the parent's latest logged request selection, or its creation options before the first request. An inherited effort survives only while the effective provider/model route is unchanged; changing the route without an explicit effort lets the selected model supply its default. A configured provider, model, or effort is checked through the optional `ctx.llm` service before child creation even when the call omits model-selection fields; a missing service or invalid value rejects the call. | | `persona` | Per-child persona; requires provider `persona` capability. | | `toolFilter` | Per-child global-tool restriction; requires `toolFilter` capability. | | `maxDepth` | Absolute delegation-depth cap, default `3` (`0` forbids delegation); a numeric cap requires the `depthLimit` capability and fails the mount without it. `'provider-managed'` sends no cap for an out-of-process provider whose budget belongs to the child harness. The tool stays visible at the cap; each attempted start checks the calling agent's current depth and returns an errored tool result when rejected. | @@ -37,15 +43,29 @@ Foreground and background calls are concurrency-safe: sibling delegations in one #### What the model sees -The generated default [`subagent` schema](../../../docs/tool-catalog.md#deepseek-aidsh-tool-subagent) under this instance's configured name while its provider exists. Provider context inheritance changes the tool and prompt descriptions. Enabled background mode adds `run_in_background`: continuable mode documents its `true` default, runtime settlement notice, and explicit foreground override, while one-shot mode documents its `false` default and the job id collected with `job_output` or stopped with `job_kill`. While the tool is visible in an assembly's scope, a `tool:` system-prompt section tells the model to start independent continuable delegations together, keep working while they run, and choose foreground only when its next action depends on the result; a tool restriction removes both its schema and this guidance. +The generated default [`subagent` schema](../../../docs/tool-catalog.md#deepseek-aidsh-tool-subagent) under this instance's configured name while its provider exists. `enableModelSelection` adds `provider`, `model`, and `reasoning_effort` plus inheritance and selection guidance; the provider must support `agentOptions`. Provider context inheritance changes the tool and prompt descriptions. Enabled background mode adds `run_in_background`: continuable mode documents its `true` default, runtime settlement notice, and explicit foreground override, while one-shot mode documents its `false` default and the job id collected with `job_output` or stopped with `job_kill`. While the tool is visible in an assembly's scope, a `tool:` system-prompt section tells the model to start independent continuable delegations together, keep working while they run, and choose foreground only when its next action depends on the result; a tool restriction removes both its schema and this guidance. #### Token effect -Fixed schema cost per parent request; each provider instance adds one schema, and each continuable instance adds one short system-prompt section. +Fixed schema cost per parent request; enabling model selection adds three parameters. Each subagent provider instance adds one schema, and each continuable instance adds one short system-prompt section. #### KV Cache effect -Prefix-stable while provider instances, names, descriptions, and schemas are unchanged. Provider registration lifecycle may invalidate parent reuse from the first changed tool definition. +Prefix-stable while subagent provider instances and their configuration are unchanged. Adapter catalog changes do not alter the definition. A route override on an inheritance-capable instance may prevent the child from reusing the inherited parent prefix. + +### Model selection and discovery + +#### What the model sees + +An instance with static `enableModelSelection: true`, or a settings-controlled instance whose Session decision is enabled, exposes the child LLM selection fields and `list_subagent_models`. Calls reject while the optional `ctx.llm` service is unavailable. With no arguments the discovery tool returns registered provider ids and names; with `provider` it returns that adapter's advertised models; with `provider` and `model` it resolves the exact model and returns its advertised reasoning efforts and default. The result is read-only runtime metadata, not an authorization list. + +#### Token effect + +One fixed tool schema is present in shipped compositions. Directory contents enter the transcript only when the model calls the tool. + +#### KV Cache effect + +The schema is prefix-stable across adapter registration and catalog changes. Each result is appended after the reusable prefix. ### Foreground result @@ -79,4 +99,5 @@ Append-only; newly visible content follows the reusable request prefix and does - **Background runs expose no result through this tool** — a one-shot task's final output is collected through the generic task surface, and a continuable child's output stays in its own session, read by its subagent id. The settlement notice states how that child ended and carries any final assistant message, but it is not this call's return value and cannot be awaited here. - **Duplicate names across waiting one-shot instances are detected late** (`TODO(subagent-dup-toolname)`) — continuable instances reserve their prompt-section name during plugin application, but preventing provider-registration rollback for waiting one-shot instances requires a registry of intended names. -- **Child policy is fixed per instance** — another model, persona, tool filter, or depth cap requires another distinctly named tool. +- **Shipped fork tools cannot select a child LLM route** — they inherit the parent's provider and model to keep the copied conversation prefix eligible for KV Cache reuse. Re-enable the fields only when route changes preserve reuse or expose a bounded recomputation cost. +- **Non-routing child policy is fixed per instance** — another persona, tool filter, or depth cap requires another distinctly named tool. LLM provider/model/reasoning-effort selection requires static enablement or an enabled per-Session preference and a subagent provider that advertises `agentOptions`; out-of-process providers currently reject enabling it rather than ignore it. diff --git a/packages/subagent/tool-subagent/README.zh.md b/packages/subagent/tool-subagent/README.zh.md index 74e8896a15..aa6dec73c6 100644 --- a/packages/subagent/tool-subagent/README.zh.md +++ b/packages/subagent/tool-subagent/README.zh.md @@ -6,7 +6,11 @@ ## 提供方选择与生命周期 -每个插件实例把一个 `provider` 绑定到一个 `toolName`；模型不会收到提供方选择器。如需公开另一种传输，请加载另一个名称不同的实例。工具只在其提供方存在时注册，从而避免对同级加载顺序和提供方重新加载的依赖。工具描述遵循 `provider.inheritsParentContext`：新建子 agent（智能体）需要独立提示词，而 fork 子 agent 已能看到父级已完成轮次。 +每个插件实例把一个 subagent 传输 `provider` 绑定到一个 `toolName`；模型不能改变该传输。如需公开另一种传输，请加载另一个名称不同的实例。`enableModelSelection: true`，或 `modelSelectionSettings: true` 时已启用的 Host 偏好，都要求该提供方具备子级 `agentOptions` 能力，并且无需额外路由配置即可公开可选的子 agent LLM `provider`、`model` 与 `reasoning_effort` 字段。调用可以提供完整的提供方／模型对；当配置值或父 Agent 值能够提供生效路由时，也可以只提供推理强度。实时 adapter 会在创建子 agent 前解析显式或配置的路由。完全省略选择字段的调用使用 `agentOptions`，再从父 Agent 最新记录的请求选择中继承兼容的缺失值；首个请求之前回退到其创建选项，并保留其中配置的 `maxTokens`。如果更换提供方或模型但没有指定强度，则清除下层路由所属的强度，使所选模型解析自己的默认值。 + +委派工具只在其 subagent 提供方存在时注册，从而避免对同级加载顺序和提供方重新加载的依赖。启用模型选择时，即使没有 `ctx.llm`，可选字段仍然可见；选择路由的调用会在该服务缺失时失败。禁用时，schema 会省略这些字段，执行阶段也会拒绝强制传入的选择。配置的 `agentOptions` 仍是部署方所有的子级默认值，不受这个面向模型的开关影响。adapter 目录和拓扑变化不会改写或重新注册工具。工具描述遵循 `provider.inheritsParentContext`：新建子 agent（智能体）需要独立提示词，而 fork 子 agent 已能看到父级已完成轮次。 + +启用的定义会注册 `list_subagent_models`，它会在调用时列出已注册提供方、某个提供方公布的模型，或某个精确模型的推理强度。因为发现工具使用全局名称，一个工具作用域最多只能由一个实例启用选择；多个持有方会使注册失败。随附产品组合默认关闭主 `subagent`（`spawn`）实例，并在每个新的顶层会话完成组合时读取 Host 的 `subagent-model-selection.enabled` 偏好。启用决定记录为 `subagent/model-selection-enabled`，由子会话继承并在恢复时保留；之后修改设置不会改变运行中的会话。组合会刻意在 `subagent_fork` 上保持禁用，使 fork 继承父级的提供方与模型：更改该路由会失去继承对话前缀的提供方侧 KV Cache 复用，重新计算前缀的成本可能超过委派任务本身。即使分离发现工具的持有权，该限制也仍然成立。目录条目仍只提供建议：如果适配器接受未列出的模型 ID，启用选择的委派工具也会接受。理由与重新开放条件由[模型选择路由 Agent Note](../../../.agents/notes/implemented/feature/2026-08-18-model-selected-subagent-routes.zh.md)负责。 前台调用会让执行信号贯穿启动和执行，等待 `run.result`，并且在返回前总会等待 `run.dispose()`。只有 `completed` 会返回规范值 `{ kind: 'foreground', runId, output: JsonValue[] }`，并渲染为相同的最终文本。中止、拒绝、token 上限和其他失败都会变成出错的工具结果，其消息依次包含终止原因标题、可选的提供方 `SubagentResult.diagnostic`，以及子 agent 保留下来的部分 assistant 文本。诊断与 `SubagentResult.output` 保持分离，因此被截断的回答不会被报告为成功，也不会与基础设施说明混淆。如果结果收集与 dispose（资源释放）都 reject，出错结果会保留两项失败。 @@ -20,9 +24,11 @@ |---|---| | `provider`（必填） | 提供方名称（`spawn`、`fork`、`acp` 等）。 | | `toolName` | 面向模型的名称，默认 `subagent`；每个已加载实例必须不同。 | +| `enableModelSelection` | 公开并接受面向模型的子级 LLM 选择字段，同时注册共享的 `list_subagent_models` 工具；默认为 `false`。它要求 subagent 提供方具备 `agentOptions` 能力。一个工具作用域最多只能由一个实例启用；即使没有 `ctx.llm`，发现 schema 仍保持注册，而发现调用和所选路由调用会在该可选服务可用前失败。禁用此开关时仍可配置 `agentOptions`。 | +| `modelSelectionSettings` | 组合 Agent 时读取 Host 的 `subagent-model-selection` 偏好，把启用决定记录进其 Session，并让子 Session 继承该决定。默认为 `false`；与 `enableModelSelection` 互斥，且只能用于 Agent 作用域组合。该偏好默认关闭，只影响之后组合的新顶层 Session。 | | `enableRunInBackground` | 公开后台模式，默认 `true`；禁用时也会拒绝强制后台调用。 | | `backgroundMode` | 后台生命周期策略，默认 `one-shot`。`one-shot` 默认前台调用；`continuable` 默认后台调用，要求提供方具备 `prepareContinuable` 能力，并返回持久化子 agent ID，且不要求加载后续消息工具。 | -| `agentOptions` | 传给具体提供方的子 agent `provider`、`model` 和正整数 `maxTokens`；进程内提供方会用显式值覆盖继承的父级选项。 | +| `agentOptions` | 配置的子 agent LLM `provider`、`model`、adapter 自有 `reasoningEffort` 与正整数 `maxTokens`；要求 subagent 提供方具备 `agentOptions` 能力。进程内提供方把显式值合并到父 Agent 最新记录的请求选择之上；首个请求之前则合并到其创建选项之上。只有生效提供方／模型路由不变时才会保留继承的推理强度；改变路由但不显式提供强度时，由所选模型提供默认值。即使调用省略模型选择字段，配置的提供方、模型或强度也会在创建子 agent 前通过可选 `ctx.llm` 服务进行校验；服务缺失或值无效都会拒绝调用。 | | `persona` | 每个子 agent 独立的 persona；要求提供方具备 `persona` 能力。 | | `toolFilter` | 每个子 agent 独立的全局工具限制；要求提供方具备 `toolFilter` 能力。 | | `maxDepth` | 绝对委派深度上限，默认 `3`（`0` 禁止委派）；数值上限要求 `depthLimit` 能力，缺失时挂载失败。对于预算由子 harness 拥有的进程外提供方，`'provider-managed'` 不发送上限。工具在达到上限时仍然可见；每次尝试启动都会检查调用 agent 的当前深度，被拒绝时返回出错的工具结果。 | @@ -37,15 +43,29 @@ #### 模型看到的内容 -当提供方存在时，以当前实例配置的名称公开已生成的默认 [`subagent` schema](../../../docs/tool-catalog.zh.md#deepseek-aidsh-tool-subagent)。提供方是否继承上下文会改变工具描述和提示词描述。启用后台模式会添加 `run_in_background`：可继续模式会记录其默认值为 `true`、运行时结算通知与显式前台覆盖；一次性模式会记录其默认值为 `false`，以及用 `job_output` 收集或用 `job_kill` 停止的 job id。当工具在本次组装的作用域中可见时，一个 `tool:` 系统提示词 section 会指示模型同时启动相互独立的可继续委派、在它们运行时继续工作，并且仅当下一步动作依赖结果时选择前台；工具限制会同时移除其 schema 和这段指引。 +当提供方存在时，以当前实例配置的名称公开已生成的默认 [`subagent` schema](../../../docs/tool-catalog.zh.md#deepseek-aidsh-tool-subagent)。`enableModelSelection` 会添加 `provider`、`model` 与 `reasoning_effort`，以及继承和选择指引；提供方必须支持 `agentOptions`。提供方是否继承上下文会改变工具描述和提示词描述。启用后台模式会添加 `run_in_background`：可继续模式会记录其默认值为 `true`、运行时结算通知与显式前台覆盖；一次性模式会记录其默认值为 `false`，以及用 `job_output` 收集或用 `job_kill` 停止的 job id。当工具在本次组装的作用域中可见时，一个 `tool:` 系统提示词 section 会指示模型同时启动相互独立的可继续委派、在它们运行时继续工作，并且仅当下一步动作依赖结果时选择前台；工具限制会同时移除其 schema 和这段指引。 #### Token 影响 -每个父级请求都会产生固定的 schema token 开销；每个提供方实例增加一个 schema，每个可继续实例还会增加一个简短的系统提示词 section。 +每个父级请求都会产生固定的 schema token 开销；启用模型选择会增加三个参数。每个 subagent 提供方实例增加一个 schema，每个可继续实例还会增加一个简短的系统提示词 section。 #### KV Cache 影响 -只要提供方实例、名称、描述和 schema 不变，前缀就保持稳定。提供方注册生命周期可能从首个变化的工具定义开始，使父级复用失效。 +只要 subagent 提供方实例及其配置不变，前缀就保持稳定。adapter 目录变化不会改变定义。具备继承能力的实例如果覆盖路由，可能阻止子 agent 复用继承的父级前缀。 + +### 模型选择与发现 + +#### 模型看到的内容 + +静态配置 `enableModelSelection: true` 的实例，或 Session 决定为启用的 settings 控制实例，会公开子级 LLM 选择字段与 `list_subagent_models`。可选 `ctx.llm` 服务不可用时，调用会失败。无参数调用发现工具会返回已注册提供方的 ID 和名称；提供 `provider` 时返回该适配器公布的模型；同时提供 `provider` 和 `model` 时解析精确模型，并返回其公布的推理强度和默认值。结果是只读的运行时元数据，不是授权列表。 + +#### Token 影响 + +随附组合会包含一个固定工具 schema。只有模型调用该工具时，目录内容才会进入 transcript。 + +#### KV Cache 影响 + +adapter 注册和目录变化不会改变 schema 的前缀稳定性。每次结果都追加在可复用前缀之后。 ### 前台结果 @@ -79,4 +99,5 @@ - **后台运行不通过本工具公开结果**：一次性任务的最终输出通过通用 Task 接口收集，可继续子 agent 的输出留在其自身会话中，按其 subagent id 读取。结算通知会说明该子 agent 如何结束，并携带可能存在的最终 assistant 消息，但它不是本次调用的返回值，也无法在此等待。 - **等待中的一次性实例较晚才发现重复名称**（`TODO(subagent-dup-toolname)`）：可继续实例会在插件应用期间预留提示词 section 名称，但若要阻止等待中的一次性实例回滚提供方注册，仍需要一份预期名称注册表。 -- **每个实例的子 agent 策略固定**：其他模型、persona、工具过滤器或深度上限都需要另一个名称不同的工具。 +- **随附 fork 工具无法选择子级 LLM 路由**：它们会继承父级的提供方与模型，使复制的对话前缀仍可供 KV Cache 复用。只有在路由变化仍能保留复用，或接口能公开一项有界的重算成本时，才重新启用这些字段。 +- **每个实例的非路由子 agent 策略固定**：其他 persona、工具过滤器或深度上限都需要另一个名称不同的工具。LLM 提供方／模型／推理强度选择要求静态启用或每 Session 偏好已启用，并要求 subagent 提供方声明 `agentOptions`；进程外提供方目前会拒绝启用它，而不是忽略它。 diff --git a/packages/subagent/tool-subagent/package.json b/packages/subagent/tool-subagent/package.json index 9ddab59996..708ca2335b 100644 --- a/packages/subagent/tool-subagent/package.json +++ b/packages/subagent/tool-subagent/package.json @@ -18,6 +18,10 @@ "types": "./lib/types/index.d.ts", "default": "./lib/index.js" }, + "./model-selection-settings": { + "types": "./lib/types/model-selection-settings.d.ts", + "default": "./lib/model-selection-settings.js" + }, "./invariant": { "types": "./lib/types/invariant.d.ts", "default": "./lib/invariant.js" @@ -28,6 +32,7 @@ "files": [ "lib/index.js", "lib/invariant.js", + "lib/model-selection-settings.js", "lib/types/**/*.d.ts" ], "license": "MIT", @@ -35,6 +40,9 @@ "@deepseek-ai/dsh-agent": "workspace:^", "@deepseek-ai/dsh-invariants": "workspace:^", "@deepseek-ai/dsh-llm": "workspace:^", + "@deepseek-ai/dsh-session": "workspace:^", + "@deepseek-ai/dsh-scope": "workspace:^", + "@deepseek-ai/dsh-settings": "workspace:^", "@deepseek-ai/dsh-subagent": "workspace:^", "@deepseek-ai/dsh-system-prompt": "workspace:^", "@deepseek-ai/dsh-jobs": "workspace:^", @@ -50,8 +58,10 @@ "@deepseek-ai/dsh-invariants": "workspace:^", "@deepseek-ai/dsh-llm": "workspace:^", "@deepseek-ai/dsh-session": "workspace:^", + "@deepseek-ai/dsh-scope": "workspace:^", "@deepseek-ai/dsh-session-persistence": "workspace:^", "@deepseek-ai/dsh-session-persistence-jsonl": "workspace:^", + "@deepseek-ai/dsh-settings": "workspace:^", "@deepseek-ai/dsh-subagent": "workspace:^", "@deepseek-ai/dsh-subagent-spawn-in-process": "workspace:^", "@deepseek-ai/dsh-system-prompt": "workspace:^", diff --git a/packages/subagent/tool-subagent/src/index.ts b/packages/subagent/tool-subagent/src/index.ts index ef0e9941eb..ceb04cced8 100644 --- a/packages/subagent/tool-subagent/src/index.ts +++ b/packages/subagent/tool-subagent/src/index.ts @@ -10,14 +10,33 @@ import type { Context } from '@deepseek-ai/cordis' import z from '@deepseek-ai/schemastery' +import { scopeChainOf, scopeOf } from '@deepseek-ai/dsh-scope' import { defineTool } from '@deepseek-ai/dsh-tools' -import type { AgentOptions } from '@deepseek-ai/dsh-agent' +import type { Agent, AgentOptions } from '@deepseek-ai/dsh-agent' +import { ReasoningEffortId } from '@deepseek-ai/dsh-llm' import type { ContentBlock } from '@deepseek-ai/dsh-llm' import type { JsonValue } from '@deepseek-ai/dsh-session' -import { assertSubagentMaxDepth, settleRun } from '@deepseek-ai/dsh-subagent' +import { + assertSubagentMaxDepth, + parentAgentOptionsForDelegation, + settleRun, +} from '@deepseek-ai/dsh-subagent' import type { SubagentProvider, SubagentResult, SubagentRun } from '@deepseek-ai/dsh-subagent' import type { JobOutcome } from '@deepseek-ai/dsh-jobs' import type {} from '@deepseek-ai/dsh-system-prompt' +import { + hasConfiguredLlmSelection, + hasDelegationModelRequest, + preflightChildLlmRoute, + requestedAgentOptions, +} from './model-selection.ts' +import type { DelegationModelRequest } from './model-selection.ts' +import { registerListSubagentModels } from './list-models.ts' +import type {} from './model-selection-settings.ts' +import { + hasSubagentModelSelection, + recordSubagentModelSelection, +} from './model-selection-state.ts' export const name = 'tool-subagent' export const inject = ['tools', 'subagents', 'systemPrompt'] @@ -34,6 +53,14 @@ export interface Config { * a distinct name. */ toolName?: string + /** Let the model discover and select the child LLM route (default false). */ + enableModelSelection?: boolean + /** + * Sample the Host `subagent-model-selection` user setting for each new + * top-level session and inherit that decision in its child sessions. Mutually + * exclusive with `enableModelSelection`. + */ + modelSelectionSettings?: boolean /** * Expose `run_in_background` (default true). Disabled instances omit the * parameter and reject forced background calls. @@ -81,14 +108,22 @@ export interface Config { export const Config: z = z.object({ provider: z.string().required(), toolName: z.string().default('subagent'), + enableModelSelection: z.boolean().default(false), + modelSelectionSettings: z.boolean().default(false), enableRunInBackground: z.boolean().default(true), backgroundMode: z.union(['one-shot', 'continuable'] as const).default('one-shot'), // Prevent Schemastery from materializing omitted agentOptions as `{}`. agentOptions: z.object({ provider: z.string(), model: z.string(), + reasoningEffort: z.string().min(1) as z>, maxTokens: z.number().step(1).min(1).max(Number.MAX_SAFE_INTEGER), - }).default(undefined as unknown as { provider: string; model: string; maxTokens: number }), + }).default(undefined as unknown as { + provider: string + model: string + reasoningEffort: ReturnType + maxTokens: number + }), persona: z.string(), // Preserve omission; Schemastery's `{ allow: [] }` default would deny every tool. toolFilter: z.object({ @@ -281,196 +316,348 @@ export function apply(ctx: Context, config: Config): void { if (config.toolFilter !== undefined && config.toolFilter.allow === undefined && config.toolFilter.deny === undefined) { throw new Error('tool-subagent: `toolFilter` is configured but names neither `allow` nor `deny` — remove the key or fill the filter') } + if (config.enableModelSelection === true && config.modelSelectionSettings === true) { + throw new Error('tool-subagent: `enableModelSelection` and `modelSelectionSettings` are mutually exclusive') + } const backgroundEnabled = config.enableRunInBackground !== false const continuable = (config.backgroundMode ?? 'one-shot') === 'continuable' const toolName = config.toolName ?? 'subagent' - // Load order and HMR replacement can change provider availability while - // this fiber remains active. - let disposeTool: (() => void) | undefined - const mount = (provider: SubagentProvider): void => { - // A numeric cap the provider cannot enforce is a misconfiguration — fail at - // mount (the earliest point the provider's capabilities are known), not on - // the first delegation. - if (typeof config.maxDepth === 'number' && !provider.capabilities.depthLimit) { + + const modelSelectionCapable = config.enableModelSelection === true || config.modelSelectionSettings === true + + const assertSubagentProviderConfiguration = (subagentProvider: SubagentProvider): void => { + if (typeof config.maxDepth === 'number' && !subagentProvider.capabilities.depthLimit) { throw new Error( - `tool-subagent: provider "${provider.name}" cannot enforce maxDepth (no depthLimit capability) — ` + `tool-subagent: provider "${subagentProvider.name}" cannot enforce maxDepth (no depthLimit capability) — ` + 'set maxDepth: \'provider-managed\' to leave the recursion budget to the provider', ) } - const wording = providerWording(provider.inheritsParentContext) - if (continuable && provider.prepareContinuable === undefined) { + if (config.agentOptions !== undefined && !subagentProvider.capabilities.agentOptions) { throw new Error( - `tool-subagent: provider "${provider.name}" does not support \`backgroundMode: continuable\``, + `tool-subagent: provider "${subagentProvider.name}" does not support child agentOptions`, + ) + } + if (modelSelectionCapable && !subagentProvider.capabilities.agentOptions) { + throw new Error( + `tool-subagent: provider "${subagentProvider.name}" does not support child model selection`, + ) + } + if (continuable && subagentProvider.prepareContinuable === undefined) { + throw new Error( + `tool-subagent: provider "${subagentProvider.name}" does not support \`backgroundMode: continuable\``, ) } - disposeTool = ctx.tools.register(defineTool({ - name: toolName, - description: wording.description + (backgroundEnabled - // The completion notice is the continuation service's own behavior, not - // a separately installed capability, so this promise holds whenever the - // continuable background path is reachable at all. - ? continuable - ? ' This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.' - : ' This call waits for the result by default. Set `run_in_background: true` to return a job id; collect with `job_output` and stop with `job_kill`.' - : ' This call waits for the subagent and returns its result.'), - parameters: { - description: { - type: 'string', - required: true, - description: 'A short (3-5 word) description of the delegated task, for display.', - }, - prompt: { - type: 'string', - required: true, - description: wording.promptDescription, - }, - ...backgroundEnabled ? { - run_in_background: { - type: 'boolean' as const, - description: continuable - ? 'Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it.' - : 'Whether to run as a background job and return its id. Defaults to false; collect with job_output or stop with job_kill.', - }, - } : {}, - }, - output: { - schema: { - oneOf: [ - { - type: 'object', - additionalProperties: false, - properties: { - kind: { type: 'string', required: true, const: 'background' }, - jobId: { type: 'string', required: true }, - }, - }, - { - type: 'object', - additionalProperties: false, - properties: { - kind: { type: 'string', required: true, const: 'continuable' }, - subagentId: { type: 'string', required: true }, - }, - }, - { - type: 'object', - additionalProperties: false, - properties: { - kind: { type: 'string', required: true, const: 'foreground' }, - runId: { type: 'string', required: true }, - output: { type: 'array', required: true, items: { type: 'json' } }, - }, - }, - ], - }, - render: (_args, value) => [{ - type: 'text', - text: value.kind === 'background' - ? `started background subagent job ${value.jobId}` - : value.kind === 'continuable' - ? `started subagent ${value.subagentId}` - : outputValueText(value.output), - }], - }, - // Children never mutate the parent session; the one parent-owned write - // (tasks.start) is a synchronous commutative insertion. - isConcurrencySafe: () => true, - async execute(args, exec) { - const parent = exec.agent - if (!parent) { - // Non-agent callers provide no parent for delegation ownership. - throw new Error('subagent tool requires a calling agent (exec.agent was undefined)') - } - - const maxDepth = typeof config.maxDepth === 'number' ? config.maxDepth : undefined - const request = { - label: args.description, - prompt: [{ type: 'text', text: args.prompt }] as ContentBlock[], - parent, - ...config.agentOptions !== undefined ? { agentOptions: config.agentOptions } : {}, - ...config.persona !== undefined ? { persona: config.persona } : {}, - ...config.toolFilter !== undefined ? { toolFilter: config.toolFilter } : {}, - ...maxDepth !== undefined ? { maxDepth } : {}, - } - - const runSpec = resolveDelegationRun(args, { backgroundEnabled, continuable }) - if (runSpec.runInBackground) { - if (continuable) { - // Resolves at inbox acceptance: the child owns its own turns from - // there, so this call neither waits for nor collects a result. - const started = await ctx.subagents.startContinuable({ - provider: config.provider, - label: args.description, - request, - signal: exec.signal, - }) - return { kind: 'continuable' as const, subagentId: started.childId } - } - const jobs = ctx.get('jobs') - if (jobs === undefined) { - throw new Error('background jobs unavailable: load @deepseek-ai/dsh-jobs and @deepseek-ai/dsh-tool-jobs') - } - // One-shot background child: job preflight finishes before the - // starter can spawn, and the task-owned signal covers startup. - const id = jobs.start({ - kind: 'subagent', - label: args.description, - owner: parent, - run: () => { - const controller = new AbortController() - const start = ctx.subagents.start(config.provider, { ...request, signal: controller.signal }) - return { - cancel: (reason?: string) => { - controller.abort(reason ?? 'background subagent task killed') - }, - done: settleStart(start, controller.signal), - // No readOutput: the child session owns intermediate detail. - } - }, - }) - return { kind: 'background' as const, jobId: id } - } - - const run: SubagentRun = await ctx.subagents.start(config.provider, { - ...request, - signal: exec.signal, - }) - return settleForegroundRun(run) - }, - })) } - // Register listeners before checking presence so no synchronous change is missed. - // TODO(subagent-dup-toolname): two waiting one-shot fibers configured with the - // same toolName collide when their provider appears, and the duplicate-name - // throw rolls back the provider registration. Continuable instances reserve - // their prompt-section name during apply() and fail earlier. Add an intent - // registry if the late one-shot collision occurs in a shipped composition. - ctx.on('subagent/provider-added', (provider) => { - if (provider.name === config.provider && disposeTool === undefined) mount(provider) + // Validate provider-owned config outside the optional LLM binding so an + // invalid provider always rejects its registration or this plugin's load. + ctx.on('subagent/provider-added', (subagentProvider) => { + if (subagentProvider.name === config.provider) assertSubagentProviderConfiguration(subagentProvider) }) - ctx.on('subagent/provider-removed', (name) => { - if (name !== config.provider || disposeTool === undefined) return - disposeTool() - disposeTool = undefined - }) - const present = ctx.subagents.getProvider(config.provider) - if (present !== undefined) { - mount(present) - } else { - // A backend fiber may activate later; a misspelled provider remains visible in this log. - ctx.logger.info(`subagent provider "${config.provider}" not registered yet; the "${config.toolName ?? 'subagent'}" tool will register when it appears`) - } - if (backgroundEnabled && continuable) { - // The section follows provider availability without its own manual - // lifecycle: empty text is omitted from rendered prompts while the tool is - // absent, and the registration itself stays owned by this plugin fiber. - ctx.systemPrompt.section({ - name: `tool:${toolName}`, - order: SUBAGENT_SECTION_ORDER, - text: context => disposeTool === undefined || ctx.tools.get(toolName, context.scope) === undefined + const initialProvider = ctx.subagents.getProvider(config.provider) + if (initialProvider !== undefined) assertSubagentProviderConfiguration(initialProvider) + + const install = (runtimeCtx: Context, modelSelectionEnabled: boolean): void => { + if (modelSelectionEnabled) registerListSubagentModels(runtimeCtx) + // Load order and HMR replacement can change provider availability while + // this fiber remains active. + let mounted: { subagentProvider: SubagentProvider; disposeTool: () => void } | undefined + const mount = (subagentProvider: SubagentProvider): void => { + assertSubagentProviderConfiguration(subagentProvider) + const wording = providerWording(subagentProvider.inheritsParentContext) + const choiceDescription = !modelSelectionEnabled ? '' - : `Use ${toolName} in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set \`run_in_background: false\` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.`, + : ' Child LLM selection is optional. Omit `provider`, `model`, and `reasoning_effort` to use configured child defaults and inherit compatible missing values from the parent Agent. Supply `provider` and `model` together after using `list_subagent_models` to inspect advertised routes and efforts. Changing the effective route without naming an effort uses the selected model\'s default effort.' + + (subagentProvider.inheritsParentContext + ? ' Changing the route can prevent provider-side reuse of the inherited conversation prefix.' + : '') + const disposeTool = runtimeCtx.tools.register(defineTool({ + name: toolName, + description: wording.description + (backgroundEnabled + // The completion notice is the continuation service's own behavior, not + // a separately installed capability, so this promise holds whenever the + // continuable background path is reachable at all. + ? continuable + ? ' This tool runs in the background by default, immediately returns a durable subagent id, and keeps the child conversation available for later turns. When that run settles, the runtime sends the parent a notice containing its outcome and any final assistant message; `send_message` starts a later turn in the same child conversation. Set `run_in_background: false` only when your next action depends on receiving the result.' + : ' This call waits for the result by default. Set `run_in_background: true` to return a job id; collect with `job_output` and stop with `job_kill`.' + : ' This call waits for the subagent and returns its result.') + choiceDescription, + parameters: { + description: { + type: 'string', + required: true, + description: 'A short (3-5 word) description of the delegated task, for display.', + }, + prompt: { + type: 'string', + required: true, + description: wording.promptDescription, + }, + ...modelSelectionEnabled ? { + provider: { + type: 'string' as const, + description: 'LLM provider route for the child. Supply together with model; omit both to use configured child defaults or inherit the parent route.', + }, + model: { + type: 'string' as const, + description: 'Model id interpreted by provider. Supply together with provider; omit both to use configured child defaults or inherit the parent route.', + }, + reasoning_effort: { + type: 'string' as const, + description: 'Adapter-owned reasoning effort for the effective child route. Omit to inherit a compatible configured/parent effort or use a newly selected model\'s default.', + }, + } : {}, + ...backgroundEnabled ? { + run_in_background: { + type: 'boolean' as const, + description: continuable + ? 'Whether to run in the background and return a durable subagent id immediately. Defaults to true. Set false to wait for the result when your next action depends on it.' + : 'Whether to run as a background job and return its id. Defaults to false; collect with job_output or stop with job_kill.', + }, + } : {}, + }, + output: { + schema: { + oneOf: [ + { + type: 'object', + additionalProperties: false, + properties: { + kind: { type: 'string', required: true, const: 'background' }, + jobId: { type: 'string', required: true }, + }, + }, + { + type: 'object', + additionalProperties: false, + properties: { + kind: { type: 'string', required: true, const: 'continuable' }, + subagentId: { type: 'string', required: true }, + }, + }, + { + type: 'object', + additionalProperties: false, + properties: { + kind: { type: 'string', required: true, const: 'foreground' }, + runId: { type: 'string', required: true }, + output: { type: 'array', required: true, items: { type: 'json' } }, + }, + }, + ], + }, + render: (_args, value) => [{ + type: 'text', + text: value.kind === 'background' + ? `started background subagent job ${value.jobId}` + : value.kind === 'continuable' + ? `started subagent ${value.subagentId}` + : outputValueText(value.output), + }], + }, + // Children never mutate the parent session; the one parent-owned write + // (tasks.start) is a synchronous commutative insertion. + isConcurrencySafe: () => true, + async execute(args, exec) { + const parent = exec.agent + if (!parent) { + // Non-agent callers provide no parent for delegation ownership. + throw new Error('subagent tool requires a calling agent (exec.agent was undefined)') + } + + const modelRequest = args as DelegationModelRequest + const parentOptions = parentAgentOptionsForDelegation(parent) + const childAgentOptions = requestedAgentOptions( + parentOptions, + config.agentOptions, + modelRequest, + modelSelectionEnabled, + ) + if (hasDelegationModelRequest(modelRequest) || hasConfiguredLlmSelection(config.agentOptions)) { + const llm = runtimeCtx.get('llm') + if (llm === undefined) { + throw new Error('cannot resolve the selected child LLM route because the `llm` service is unavailable') + } + await preflightChildLlmRoute(llm, parentOptions, childAgentOptions, exec.signal) + } + exec.signal.throwIfAborted() + const maxDepth = typeof config.maxDepth === 'number' ? config.maxDepth : undefined + const request = { + label: args.description, + prompt: [{ type: 'text', text: args.prompt }] as ContentBlock[], + parent, + ...childAgentOptions !== undefined ? { agentOptions: childAgentOptions } : {}, + ...config.persona !== undefined ? { persona: config.persona } : {}, + ...config.toolFilter !== undefined ? { toolFilter: config.toolFilter } : {}, + ...maxDepth !== undefined ? { maxDepth } : {}, + } + + const runSpec = resolveDelegationRun(args, { backgroundEnabled, continuable }) + if (runSpec.runInBackground) { + if (continuable) { + // Resolves at inbox acceptance: the child owns its own turns from + // there, so this call neither waits for nor collects a result. + const started = await runtimeCtx.subagents.startContinuable({ + provider: config.provider, + label: args.description, + request, + signal: exec.signal, + }) + return { kind: 'continuable' as const, subagentId: started.childId } + } + const jobs = runtimeCtx.get('jobs') + if (jobs === undefined) { + throw new Error('background jobs unavailable: load @deepseek-ai/dsh-jobs and @deepseek-ai/dsh-tool-jobs') + } + // One-shot background child: job preflight finishes before the + // starter can spawn, and the task-owned signal covers startup. + const id = jobs.start({ + kind: 'subagent', + label: args.description, + owner: parent, + run: () => { + const controller = new AbortController() + const start = runtimeCtx.subagents.start(config.provider, { ...request, signal: controller.signal }) + return { + cancel: (reason?: string) => { + controller.abort(reason ?? 'background subagent task killed') + }, + done: settleStart(start, controller.signal), + // No readOutput: the child session owns intermediate detail. + } + }, + }) + return { kind: 'background' as const, jobId: id } + } + + const run: SubagentRun = await runtimeCtx.subagents.start(config.provider, { + ...request, + signal: exec.signal, + }) + return settleForegroundRun(run) + }, + })) + mounted = { subagentProvider, disposeTool } + } + + // Register listeners before checking presence so no synchronous change is missed. + // TODO(subagent-dup-toolname): two waiting one-shot fibers configured with the + // same toolName collide when their provider appears, and the duplicate-name + // throw rolls back the provider registration. Continuable instances reserve + // their prompt-section name during apply() and fail earlier. Add an intent + // registry if the late one-shot collision occurs in a shipped composition. + runtimeCtx.on('subagent/provider-added', (subagentProvider) => { + if (subagentProvider.name === config.provider && mounted === undefined) mount(subagentProvider) + }) + runtimeCtx.on('subagent/provider-removed', (name) => { + if (name !== config.provider || mounted === undefined) return + mounted.disposeTool() + mounted = undefined + }) + const present = runtimeCtx.subagents.getProvider(config.provider) + if (present !== undefined) { + mount(present) + } else { + // A backend fiber may activate later; a misspelled provider remains visible in this log. + runtimeCtx.logger.info(`subagent provider "${config.provider}" not registered yet; the "${config.toolName ?? 'subagent'}" tool will register when it appears`) + } + if (backgroundEnabled && continuable) { + // The section follows provider availability without its own manual + // lifecycle: empty text is omitted from rendered prompts while the tool is + // absent, and the registration itself stays owned by this plugin fiber. + runtimeCtx.systemPrompt.section({ + name: `tool:${toolName}`, + order: SUBAGENT_SECTION_ORDER, + text: context => mounted === undefined || runtimeCtx.tools.get(toolName, context.scope) === undefined + ? '' + : `Use ${toolName} in the background by default. Start independent delegations together in one assistant message and continue useful work while they run. Set \`run_in_background: false\` only when your next action depends on that subagent's result. When a background run settles, the runtime sends you a notice containing its outcome and any final assistant message.`, + }) + } + } + + if (config.modelSelectionSettings !== true) { + install(ctx, config.enableModelSelection === true) + return + } + + const settings = ctx.get('subagentModelSelection') + if (settings === undefined) { + throw new Error( + 'tool-subagent: `modelSelectionSettings` requires ' + + '@deepseek-ai/dsh-tool-subagent/model-selection-settings in the Host scope', + ) + } + const compositionScope = scopeOf(ctx) + if (compositionScope === undefined) { + throw new Error('tool-subagent: `modelSelectionSettings` requires an Agent or preset scope') + } + + const selectForAgent = (agent: NonNullable): boolean => { + let enabled = hasSubagentModelSelection(agent.session) + if (!enabled) { + const parentId = agent.session.header.origin === 'subagent' + ? agent.session.header.parentSession + : undefined + if (parentId !== undefined) { + const parent = ctx.get('agents')?.get(parentId) + enabled = parent !== undefined && hasSubagentModelSelection(parent.session) + } else if (agent.session.firstLiveSeq === 0) { + enabled = settings.currentEnabled() + } + } + if (enabled) recordSubagentModelSelection(agent.session) + return enabled + } + + const agent = ctx.agent + if (agent !== undefined) { + install(ctx, selectForAgent(agent)) + return + } + const agents = ctx.get('agents') + /* v8 ignore next -- Agent and preset scopes are minted only by the Agent registry. */ + if (agents === undefined) throw new Error('tool-subagent: scoped model-selection settings require the Agent registry') + const scopedInstalls = new WeakMap>() + const installing = new WeakSet() + const belongsToComposition = (candidate: Agent): boolean => + scopeChainOf(scopeOf(candidate.ctx)).includes(compositionScope) + const installScoped = (candidate: Agent): void => { + if (scopedInstalls.has(candidate) || installing.has(candidate)) return + // Reserve before the injected fiber runs: tool registration emits + // `tools/change` synchronously, which re-enters the reconciliation below. + installing.add(candidate) + const enabled = selectForAgent(candidate) + const fiber = candidate.ctx.inject(['tools', 'subagents', 'systemPrompt'], (runtimeCtx) => { + install(runtimeCtx, enabled) + }) + installing.delete(candidate) + scopedInstalls.set(candidate, fiber) + } + const removeScoped = (candidate: Agent): void => { + const fiber = scopedInstalls.get(candidate) + if (fiber === undefined) return + scopedInstalls.delete(candidate) + /* v8 ignore next 3 -- Cordis Fiber disposal contains registration cleanup failures; this is the final diagnostic sink. */ + void fiber.dispose().catch((error: unknown) => { + ctx.logger.warn(`tool-subagent: failed to remove recomposed Agent "${candidate.id}" definitions: ${String(error)}`) }) } + const reconcileComposedAgents = (): void => { + // Every Agent and preset scope is minted by the Agent registry; the scope + // check above makes this same-process typed relationship authoritative. + for (const candidate of agents.list()) { + if (belongsToComposition(candidate)) installScoped(candidate) + else removeScoped(candidate) + } + } + // A shipped preset is mounted once in a standing scope. Its listener admits + // only descendant Agents and installs the sampled tool definition in each + // Agent's own scope, so a later settings change cannot mutate a live session. + ctx.on('agent/created', ({ agent: created }) => { + installScoped(created) + }) + ctx.on('agent/disposed', ({ agent: disposed }) => { removeScoped(disposed) }) + // Reparenting an Agent between standing presets changes its inherited tool + // set and emits `tools/change`; reconcile the Agent-owned override with the + // new ancestry. Other registry changes are idempotent no-ops here. + ctx.on('tools/change', reconcileComposedAgents) } diff --git a/packages/subagent/tool-subagent/src/invariant.ts b/packages/subagent/tool-subagent/src/invariant.ts index 5b8facc900..84bd209caa 100644 --- a/packages/subagent/tool-subagent/src/invariant.ts +++ b/packages/subagent/tool-subagent/src/invariant.ts @@ -5,7 +5,8 @@ /* jscpd:ignore-start */ import type { Context } from '@deepseek-ai/cordis' -import type { InvariantInstaller } from '@deepseek-ai/dsh-invariants' +import type { InvariantFailure, InvariantInstaller } from '@deepseek-ai/dsh-invariants' +import { hasSubagentModelSelection } from './model-selection-state.ts' const PACKAGE_NAME = '@deepseek-ai/dsh-tool-subagent' @@ -14,11 +15,24 @@ export const name = 'tool-subagent-invariant' /** Service required before the companion can reserve package ownership. */ export const inject = ['invariants'] -/** - * No runtime invariant: this model-facing adapter has no independent lifecycle stream; execution - * relations are owned by the capability seam it calls. - */ -const install: InvariantInstaller = () => {} +/** Assert that a durable opt-in is represented by both model-facing definitions. */ +const install: InvariantInstaller = Object.assign((ctx: Context, fail: InvariantFailure) => { + ctx.on('agent/pre-step', async ({ agent }, next) => { + if (hasSubagentModelSelection(agent.session)) { + const schemas = ctx.tools.schemas(agent) + const selectable = schemas.some((schema) => { + const properties = (schema.parameters as { properties?: Record }).properties + return properties?.['provider'] !== undefined + && properties['model'] !== undefined + && properties['reasoning_effort'] !== undefined + }) + if (!selectable || !schemas.some(schema => schema.name === 'list_subagent_models')) { + fail('a subagent/model-selection-enabled session must expose route fields and list_subagent_models') + } + } + return next() + }, { global: true }) +}, { inject: ['tools'] }) /** * Register this package's invariant companion. diff --git a/packages/subagent/tool-subagent/src/list-models.ts b/packages/subagent/tool-subagent/src/list-models.ts new file mode 100644 index 0000000000..9e1ff5c24e --- /dev/null +++ b/packages/subagent/tool-subagent/src/list-models.ts @@ -0,0 +1,94 @@ +/** Model-facing discovery of LLM routes available to child Agents. */ + +import type { Context } from '@deepseek-ai/cordis' +import type LlmRuntime from '@deepseek-ai/dsh-llm' +import type { LlmProviderInfo } from '@deepseek-ai/dsh-llm' +import { defineTool } from '@deepseek-ai/dsh-tools' + +interface ListSubagentModelsRequest { + readonly provider?: string + readonly model?: string +} + +/** Resolve one registered provider with a model-correctable diagnostic. */ +function registeredProvider(llm: LlmRuntime, providerId: string): LlmProviderInfo { + const providers = llm.listProviders() + const provider = providers.find(candidate => candidate.id === providerId) + if (provider !== undefined) return provider + const available = providers.map(candidate => candidate.id).join(', ') || '(none)' + throw new Error(`LLM provider "${providerId}" is not registered; available providers: ${available}`) +} + +/** Render one advertised or resolved model. */ +function modelLine(provider: string, model: { id: string; name: string; description?: string }): string { + return `${provider}/${model.id} — ${model.name}${model.description === undefined ? '' : `: ${model.description}`}` +} + +/** Read the requested provider, advertised models, or exact-model efforts. */ +async function listSubagentModels( + ctx: Context, + request: ListSubagentModelsRequest, + signal: AbortSignal, +): Promise { + const llm = ctx.get('llm') + if (llm === undefined) { + throw new Error('cannot discover child LLM routes because the `llm` service is unavailable') + } + if (request.model !== undefined && request.provider === undefined) { + throw new Error('`model` requires `provider`') + } + if (request.provider === undefined) { + const providers = llm.listProviders() + return providers.length === 0 + ? '(no LLM providers)' + : providers.map(provider => `${provider.id} — ${provider.name}`).join('\n') + } + if (request.provider.length === 0) throw new Error('`provider` must be non-empty') + const provider = registeredProvider(llm, request.provider) + if (request.model === undefined) { + const models = await llm.listModels(provider.id) + return models.length === 0 + ? `(no advertised models for ${provider.id})` + : models.map(model => modelLine(provider.id, model)).join('\n') + } + if (request.model.length === 0) throw new Error('`model` must be non-empty') + const model = await llm.resolveModelInfo(provider.id, request.model, signal) + const efforts = model.reasoning?.efforts.map(effort => ( + `${effort.id}${model.reasoning?.defaultEffort === effort.id ? ' (default)' : ''} — ${effort.name}` + + (effort.description === undefined ? '' : `: ${effort.description}`) + )).join('\n') || '(no advertised reasoning efforts)' + return `${modelLine(provider.id, model)}\nReasoning efforts:\n${efforts}` +} + +/** + * Register `list_subagent_models` for one owning delegation-tool instance. + * @param ctx - Context whose tool registry owns the fixed discovery definition. + */ +export function registerListSubagentModels(ctx: Context): void { + ctx.tools.register(defineTool({ + name: 'list_subagent_models', + description: + 'Discover LLM routes for subagents without changing the current Agent. Call with no arguments to list ' + + 'registered providers, with `provider` to list its advertised models, or with `provider` and `model` ' + + 'to inspect that exact model and its reasoning efforts. Catalog membership is advisory: an adapter may ' + + 'accept an unlisted model id. Use the returned ids with a delegation tool\'s `provider`, `model`, and ' + + '`reasoning_effort` fields.', + parameters: { + provider: { + type: 'string', + description: 'Registered LLM provider id. Omit to list providers.', + }, + model: { + type: 'string', + description: 'Exact model id to inspect. Requires provider; omit to list that provider\'s advertised models.', + }, + }, + output: { + schema: { type: 'string' }, + render: (_args, result) => [{ type: 'text', text: result }], + }, + execute(args, exec) { + return listSubagentModels(ctx, args, exec.signal) + }, + })) +} diff --git a/packages/subagent/tool-subagent/src/model-selection-settings.ts b/packages/subagent/tool-subagent/src/model-selection-settings.ts new file mode 100644 index 0000000000..113cd4c6f8 --- /dev/null +++ b/packages/subagent/tool-subagent/src/model-selection-settings.ts @@ -0,0 +1,70 @@ +/** Host-owned opt-in setting for model-selectable subagent delegation. */ + +import { Context, Service } from '@deepseek-ai/cordis' +import z from '@deepseek-ai/schemastery' +import { installSettingsSection, settingsNamespace } from '@deepseek-ai/dsh-settings' + +declare module '@deepseek-ai/cordis' { + interface Context { + /** User preference sampled when a new Agent receives its delegation tools. */ + subagentModelSelection: SubagentModelSelectionConfig + } +} + +/** User-settings section for model-selectable subagent delegation. */ +export const SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE = settingsNamespace('subagent-model-selection') + +/** Stored user preference; the shipped composition defaults it off. */ +export interface SubagentModelSelectionSettings { + /** Whether new Agents may expose child LLM route selection to the model. */ + enabled: boolean +} + +/** Schema served to settings clients for the opt-in preference. */ +export const SUBAGENT_MODEL_SELECTION_SETTINGS_SCHEMA: z = z.object({ + enabled: z.boolean().default(false), +}) + +/** Optional deployment base for the preference. */ +export interface Config { + /** Initial value inherited when the user document does not override it. */ + enabled?: boolean +} + +/** Singleton settings owner read by delegation tools when an Agent is published. */ +export class SubagentModelSelectionConfig extends Service { + static Config: z = z.object({ + enabled: z.boolean().default(false), + }) + + private source: () => SubagentModelSelectionSettings + + constructor(ctx: Context, config: Config = {}) { + super(ctx, 'subagentModelSelection') + const entry: SubagentModelSelectionSettings = { enabled: config.enabled === true } + this.source = () => entry + installSettingsSection( + ctx, + SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, + SUBAGENT_MODEL_SELECTION_SETTINGS_SCHEMA, + entry, + { + setSource: (source) => { this.source = source }, + // Consumers sample at Agent publication, so a settings update never + // rebuilds the tool definitions of an Agent that is already running. + onChange: () => {}, + }, + ) + } + + /** + * Read the preference for the next eligible Agent publication. + * @returns whether that Agent should receive model-selectable delegation. + */ + currentEnabled(): boolean { + return this.source().enabled + } +} + +export const name = 'subagent-model-selection-settings' +export default SubagentModelSelectionConfig diff --git a/packages/subagent/tool-subagent/src/model-selection-state.ts b/packages/subagent/tool-subagent/src/model-selection-state.ts new file mode 100644 index 0000000000..35345ac115 --- /dev/null +++ b/packages/subagent/tool-subagent/src/model-selection-state.ts @@ -0,0 +1,33 @@ +/** Durable per-session state for the user-controlled model-selection opt-in. */ + +import type { Session } from '@deepseek-ai/dsh-session' + +declare module '@deepseek-ai/dsh-session/types' { + interface SessionEventMap { + /** + * Records that this session's delegation tool exposes child provider, + * model, and reasoning-effort selection. Appended before the first model + * request; absence means the fixed-route definition. Log-only: it carries + * no `surfaceOp` and never enters model history. + */ + 'subagent/model-selection-enabled': Record + } +} + +/** + * Whether a session log records the enabled model-selection definition. + * @param session - session whose durable decision is read. + * @returns whether model-selectable delegation is enabled for the session. + */ +export function hasSubagentModelSelection(session: Session): boolean { + return session.events.some(event => event.type === 'subagent/model-selection-enabled') +} + +/** + * Append the enabled decision once, before its definition can reach a model request. + * @param session - session receiving the enabled decision. + */ +export function recordSubagentModelSelection(session: Session): void { + if (hasSubagentModelSelection(session)) return + session.append('subagent/model-selection-enabled', {}) +} diff --git a/packages/subagent/tool-subagent/src/model-selection.ts b/packages/subagent/tool-subagent/src/model-selection.ts new file mode 100644 index 0000000000..6b89d92742 --- /dev/null +++ b/packages/subagent/tool-subagent/src/model-selection.ts @@ -0,0 +1,112 @@ +/** Child LLM route selection for the subagent tool. */ + +import { ReasoningEffortId } from '@deepseek-ai/dsh-llm' +import type { LlmRuntime } from '@deepseek-ai/dsh-llm' +import type { AgentOptions } from '@deepseek-ai/dsh-agent' + +/** Model-facing child LLM route fields. */ +export interface DelegationModelRequest { + readonly provider?: string + readonly model?: string + readonly reasoning_effort?: string +} + +/** + * Whether a call explicitly selects any child LLM value. + * @param request - Model-facing route fields from the tool call. + * @returns Whether at least one route or effort field is present. + */ +export function hasDelegationModelRequest(request: DelegationModelRequest): boolean { + return request.provider !== undefined + || request.model !== undefined + || request.reasoning_effort !== undefined +} + +/** Reject an empty model-facing route value at the tool JSON boundary. */ +function assertNonEmpty(value: string | undefined, field: keyof DelegationModelRequest): void { + if (value !== undefined && value.length === 0) { + throw new Error(`child LLM \`${field}\` must be non-empty`) + } +} + +/** + * Merge model-supplied selection fields over configured child defaults. + * Provider and model form one route and must be supplied together. Changing + * that route without an effort clears the configured route-owned effort. + * @param parentOptions - Current parent values that supply missing child values. + * @param configured - Tool-instance child defaults. + * @param request - Model-facing route override. + * @param enabled - Whether this tool instance permits model-facing selection. + * @returns Child Agent options, preserving omission when no layer contributes one. + */ +export function requestedAgentOptions( + parentOptions: AgentOptions, + configured: AgentOptions | undefined, + request: DelegationModelRequest, + enabled: boolean, +): AgentOptions | undefined { + if (!hasDelegationModelRequest(request)) return configured + if (!enabled) { + throw new Error('child model selection is disabled for this tool instance') + } + assertNonEmpty(request.provider, 'provider') + assertNonEmpty(request.model, 'model') + assertNonEmpty(request.reasoning_effort, 'reasoning_effort') + if ((request.provider === undefined) !== (request.model === undefined)) { + throw new Error('child LLM `provider` and `model` must be supplied together') + } + + const baselineProvider = configured?.provider ?? parentOptions.provider + const baselineModel = configured?.model ?? parentOptions.model + const routeChanged = request.provider !== undefined + && (request.provider !== baselineProvider || request.model !== baselineModel) + const { reasoningEffort: _configuredReasoningEffort, ...configuredWithoutReasoning } = configured ?? {} + return { + ...routeChanged && request.reasoning_effort === undefined ? configuredWithoutReasoning : configured, + ...request.provider === undefined ? {} : { provider: request.provider, model: request.model }, + ...request.reasoning_effort === undefined + ? {} + : { reasoningEffort: ReasoningEffortId(request.reasoning_effort) }, + } +} + +/** + * Whether configured Agent options require route validation before delegation. + * @param options - Tool-instance child defaults. + * @returns Whether configured provider, model, or effort values must be resolved. + */ +export function hasConfiguredLlmSelection(options: AgentOptions | undefined): boolean { + return options?.provider !== undefined + || options?.model !== undefined + || options?.reasoningEffort !== undefined +} + +/** + * Resolve an effective child route through its live adapter before the child is + * created. The LLM runtime owns provider lookup, exact-model metadata, effort + * validation, and adapter defaults. + * @param llm - Live LLM runtime. + * @param parentOptions - Current parent values whose compatible fields the child inherits. + * @param requested - Per-child options after request/config merging. + * @param signal - Tool-call cancellation signal. + */ +export async function preflightChildLlmRoute( + llm: LlmRuntime, + parentOptions: AgentOptions, + requested: AgentOptions | undefined, + signal: AbortSignal, +): Promise { + const provider = requested?.provider ?? parentOptions.provider + const model = requested?.model ?? parentOptions.model + if (provider === undefined || model === undefined) { + throw new Error('cannot select child LLM values without an effective provider and model') + } + const routeChanged = provider !== parentOptions.provider || model !== parentOptions.model + const reasoningEffort = requested?.reasoningEffort + ?? (routeChanged ? undefined : parentOptions.reasoningEffort) + await llm.resolveCallConfig({ + provider, + model, + ...reasoningEffort === undefined ? {} : { reasoningEffort }, + }, signal) +} diff --git a/packages/subagent/tool-subagent/tests/harness.ts b/packages/subagent/tool-subagent/tests/harness.ts new file mode 100644 index 0000000000..36ac602c89 --- /dev/null +++ b/packages/subagent/tool-subagent/tests/harness.ts @@ -0,0 +1,57 @@ +import { Context } from '@deepseek-ai/cordis' +import LlmRuntime, { CallId } from '@deepseek-ai/dsh-llm' +import SystemPrompt from '@deepseek-ai/dsh-system-prompt' +import ToolRuntime from '@deepseek-ai/dsh-tools' +import type { Agent } from '@deepseek-ai/dsh-agent' +import SubagentRuntime from '@deepseek-ai/dsh-subagent' +import { Session, SessionId } from '@deepseek-ai/dsh-session' +import * as mock from './scripted-provider.ts' +import * as tool from '../src/index.ts' + +/** Shared non-aborted tool signal for package-local integration tests. */ +export const testToolSignal = new AbortController().signal + +/** Build the minimal parent Agent owned by the package-local scripted provider. */ +export function fakeAgent(id = 'parent-1'): Agent { + const sessionId = SessionId(id) + return { id: sessionId, options: {}, session: Session.create(sessionId) } as unknown as Agent +} + +/** Mount the real tool and service stack around one scripted subagent provider. */ +export async function setup(toolConfig: tool.Config, mockConfig: Partial = {}): Promise { + const ctx = new Context() + await ctx.plugin(LlmRuntime) + await ctx.plugin(SystemPrompt) + await ctx.plugin(ToolRuntime) + await ctx.plugin(SubagentRuntime) + await mock.mountScriptedProvider(ctx, { name: 'mock', ...mockConfig }) + await ctx.plugin(tool, toolConfig) + return ctx +} + +let callCounter = 0 + +/** Execute the registered subagent tool through the real ToolRuntime pipeline. */ +export function callSubagent( + ctx: Context, + args: unknown, + over: { agent?: Agent | undefined; signal?: AbortSignal } = {}, +) { + // Distinguish "no override" (use a default agent) from an explicit + // `{ agent: undefined }` (test the no-agent path). Under + // exactOptionalPropertyTypes the key is omitted rather than set to undefined. + const agent = 'agent' in over ? over.agent : fakeAgent() + return ctx.tools.execute({ + signal: testToolSignal, + callId: CallId(`call-${++callCounter}`), + name: 'subagent', + arguments: args, + ...agent ? { agent } : {}, + ...over.signal ? { signal: over.signal } : {}, + }) +} + +/** Join text blocks from one rendered tool result. */ +export function text(result: { content: { type: string; text?: string }[] }): string { + return result.content.filter(block => block.type === 'text').map(block => block.text).join('') +} diff --git a/packages/subagent/tool-subagent/tests/list-models.spec.ts b/packages/subagent/tool-subagent/tests/list-models.spec.ts new file mode 100644 index 0000000000..de3643fd97 --- /dev/null +++ b/packages/subagent/tool-subagent/tests/list-models.spec.ts @@ -0,0 +1,191 @@ +import { describe, expect, it } from 'vitest' +import { Context } from '@deepseek-ai/cordis' +import LlmRuntime, { + CallId, + LlmAdapter, + ReasoningEffortId, +} from '@deepseek-ai/dsh-llm' +import type { + GenerateOptions, + LlmModelInfo, + LlmResolvedModelInfo, + StreamChunk, +} from '@deepseek-ai/dsh-llm' +import ToolRuntime from '@deepseek-ai/dsh-tools' +import SystemPrompt from '@deepseek-ai/dsh-system-prompt' +import SubagentRuntime from '@deepseek-ai/dsh-subagent' +import * as tool from '../src/index.ts' +import { testToolSignal, text } from './harness.ts' + +class CatalogAdapter extends LlmAdapter { + constructor(private readonly empty = false) { + super() + } + + override providerInfo(provider: string) { + return { id: provider, name: `${provider.toUpperCase()} API` } + } + + override listModels(provider: string): Promise { + if (this.empty) return Promise.resolve([]) + return Promise.resolve([ + { provider, id: 'fast', name: 'Fast', description: 'Focused work.' }, + { provider, id: 'plain', name: 'Plain' }, + ]) + } + + override resolveModel(provider: string, model: string): Promise { + if (model === 'plain') return Promise.resolve({ provider, id: model, name: 'Plain' }) + return Promise.resolve({ + provider, + id: model, + name: 'Fast', + description: 'Focused work.', + reasoning: { + efforts: [ + { id: ReasoningEffortId('low'), name: 'Low' }, + { id: ReasoningEffortId('high'), name: 'High', description: 'Quality first.' }, + ], + defaultEffort: ReasoningEffortId('high'), + }, + }) + } + + stream(_options: GenerateOptions): AsyncIterable { + return (async function* () { yield { type: 'finish' as const, reason: { kind: 'stop' as const } } })() + } +} + +async function setupListTool() { + const ctx = new Context() + await ctx.plugin(LlmRuntime) + await ctx.plugin(SystemPrompt) + await ctx.plugin(ToolRuntime) + await ctx.plugin(SubagentRuntime) + const fiber = await ctx.plugin(tool, { provider: 'unused', enableModelSelection: true }) + return { ctx, fiber } +} + +let counter = 0 + +function call(ctx: Context, args: unknown) { + return ctx.tools.execute({ + signal: testToolSignal, + callId: CallId(`list-models-${++counter}`), + name: 'list_subagent_models', + arguments: args, + }) +} + +describe('list_subagent_models', () => { + it('is omitted unless its delegation-tool instance owns discovery', async () => { + const ctx = new Context() + await ctx.plugin(LlmRuntime) + await ctx.plugin(SystemPrompt) + await ctx.plugin(ToolRuntime) + await ctx.plugin(SubagentRuntime) + await ctx.plugin(tool, { provider: 'unused' }) + expect(ctx.tools.get('list_subagent_models')).toBeUndefined() + }) + + it('stays registered without the optional LLM service and rejects discovery calls', async () => { + const ctx = new Context() + await ctx.plugin(SystemPrompt) + await ctx.plugin(ToolRuntime) + await ctx.plugin(SubagentRuntime) + await ctx.plugin(tool, { provider: 'unused', enableModelSelection: true }) + const result = await call(ctx, {}) + expect(result.isError).toBe(true) + expect(text(result)).toContain('`llm` service is unavailable') + }) + + it('rejects two discovery-owning instances in one tool scope', async () => { + const { ctx } = await setupListTool() + await expect(ctx.plugin(tool, { + provider: 'another-unused', + toolName: 'subagent_other', + enableModelSelection: true, + }).then(() => undefined)).rejects.toThrow('tool "list_subagent_models" is already registered') + }) + + it('lists registered providers and follows live registration changes', async () => { + const { ctx, fiber } = await setupListTool() + const empty = await call(ctx, {}) + expect(empty.isError).toBe(false) + expect(text(empty)).toBe('(no LLM providers)') + + const registration = ctx.llm.registerAdapter(['alpha'], new CatalogAdapter()) + const providers = await call(ctx, {}) + expect(providers.isError).toBe(false) + expect(text(providers)).toBe('alpha — ALPHA API') + + registration.replace(['beta']) + const changed = await call(ctx, {}) + expect(text(changed)).toBe('beta — BETA API') + + await fiber.dispose() + expect(ctx.tools.get('list_subagent_models')).toBeUndefined() + }) + + it('lists one provider\'s advertised models without treating the catalog as a whitelist', async () => { + const { ctx } = await setupListTool() + ctx.llm.registerAdapter(['alpha'], new CatalogAdapter()) + const result = await call(ctx, { provider: 'alpha' }) + expect(result.isError).toBe(false) + expect(text(result)).toBe('alpha/fast — Fast: Focused work.\nalpha/plain — Plain') + }) + + it('renders an empty advertised model list', async () => { + const { ctx } = await setupListTool() + ctx.llm.registerAdapter(['alpha'], new CatalogAdapter(true)) + const result = await call(ctx, { provider: 'alpha' }) + expect(result.isError).toBe(false) + expect(text(result)).toBe('(no advertised models for alpha)') + }) + + it('inspects exact-model efforts, descriptions, and defaults', async () => { + const { ctx } = await setupListTool() + ctx.llm.registerAdapter(['alpha'], new CatalogAdapter()) + const result = await call(ctx, { provider: 'alpha', model: 'fast' }) + expect(result.isError).toBe(false) + expect(text(result)).toBe( + 'alpha/fast — Fast: Focused work.\nReasoning efforts:\n' + + 'low — Low\nhigh (default) — High: Quality first.', + ) + }) + + it('renders exact models without reasoning metadata', async () => { + const { ctx } = await setupListTool() + ctx.llm.registerAdapter(['alpha'], new CatalogAdapter()) + const result = await call(ctx, { provider: 'alpha', model: 'plain' }) + expect(result.isError).toBe(false) + expect(text(result)).toBe('alpha/plain — Plain\nReasoning efforts:\n(no advertised reasoning efforts)') + }) + + it.each([ + { args: { model: 'fast' }, expected: '`model` requires `provider`' }, + { args: { provider: '' }, expected: '`provider` must be non-empty' }, + { args: { provider: 'missing' }, expected: 'available providers: (none)' }, + ])('rejects incomplete or unavailable provider requests', async ({ args, expected }) => { + const { ctx } = await setupListTool() + const result = await call(ctx, args) + expect(result.isError).toBe(true) + expect(text(result)).toContain(expected) + }) + + it('rejects an empty exact model after resolving the provider', async () => { + const { ctx } = await setupListTool() + ctx.llm.registerAdapter(['alpha'], new CatalogAdapter()) + const result = await call(ctx, { provider: 'alpha', model: '' }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('`model` must be non-empty') + }) + + it('reports registered alternatives for an unavailable provider', async () => { + const { ctx } = await setupListTool() + ctx.llm.registerAdapter(['alpha'], new CatalogAdapter()) + const result = await call(ctx, { provider: 'missing' }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('available providers: alpha') + }) +}) diff --git a/packages/subagent/tool-subagent/tests/model-selection-settings.spec.ts b/packages/subagent/tool-subagent/tests/model-selection-settings.spec.ts new file mode 100644 index 0000000000..4f95db088f --- /dev/null +++ b/packages/subagent/tool-subagent/tests/model-selection-settings.spec.ts @@ -0,0 +1,248 @@ +/** Default-off settings and per-session model-selection decisions. */ + +import { describe, expect, it, vi } from 'vitest' +import { Context } from '@deepseek-ai/cordis' +import { Session, SessionId } from '@deepseek-ai/dsh-session' +import type { SessionEvent } from '@deepseek-ai/dsh-session' +import { bindScopeParent, createScope, scopeOf, scopeTarget } from '@deepseek-ai/dsh-scope' +import { SettingsProvider } from '@deepseek-ai/dsh-settings' +import type { SettingsNamespace } from '@deepseek-ai/dsh-settings' +import InvariantRegistry from '@deepseek-ai/dsh-invariants' +import AgentLoop from '@deepseek-ai/dsh-agent-loop' +import { mountAgentLoopTestDependencies } from '@deepseek-ai/dsh-agent-loop-testkit' +import SubagentRuntime from '@deepseek-ai/dsh-subagent' +import * as SubagentSpawn from '@deepseek-ai/dsh-subagent-spawn-in-process' +import * as tool from '../src/index.ts' +import * as ToolInvariant from '../src/invariant.ts' +import SubagentModelSelectionConfig, { + SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, +} from '../src/model-selection-settings.ts' +import { hasSubagentModelSelection } from '../src/model-selection-state.ts' + +/** Writable in-memory settings provider for the package integration. */ +class MemorySettings extends SettingsProvider { + doc: Record = {} + + get writable(): boolean { + return true + } + + protected load(): Promise> { + return Promise.resolve(structuredClone(this.doc)) + } + + protected persist(ns: SettingsNamespace, section: Record): Promise { + this.doc = { ...this.doc, [ns]: structuredClone(section) } + return Promise.resolve() + } +} + +/** Read whether one Agent's delegation definition contains route fields. */ +function selectable(ctx: Context, agent: Awaited>['agent']): boolean { + const schema = ctx.tools.schemas(agent).find(candidate => candidate.name === 'subagent') + const properties = (schema?.parameters as { properties?: Record } | undefined)?.properties + return properties?.['provider'] !== undefined + && properties['model'] !== undefined + && properties['reasoning_effort'] !== undefined + && ctx.tools.schemas(agent).some(candidate => candidate.name === 'list_subagent_models') +} + +/** Mount the real settings, Agent, provider, and tool services. */ +async function boot(): Promise { + const ctx = new Context() + await ctx.plugin(MemorySettings) + await ctx.plugin(SubagentModelSelectionConfig) + await mountAgentLoopTestDependencies(ctx) + await ctx.plugin(AgentLoop, { agents: [] }) + await ctx.plugin(SubagentRuntime) + await ctx.plugin(SubagentSpawn, { providerName: 'spawn' }) + return ctx +} + +/** Create one Agent whose setup mounts the settings-controlled tool preset row. */ +async function createAgent(ctx: Context, id: string, options: { + meta?: { parentSession: SessionId; origin: 'subagent' } + seed?: readonly SessionEvent[] +} = {}) { + const handle = await ctx.agents.create({ + sessionId: SessionId(id), + ...options, + setup: async (agentCtx) => { + await agentCtx.plugin(tool, { + provider: 'spawn', + modelSelectionSettings: true, + backgroundMode: 'continuable', + }) + }, + }) + return handle.agent +} + +describe('SubagentModelSelectionConfig', () => { + it('uses the composed default without a settings provider', async () => { + const ctx = new Context() + await ctx.plugin(SubagentModelSelectionConfig, { enabled: true }) + + expect(ctx.subagentModelSelection.currentEnabled()).toBe(true) + await ctx.fiber.dispose() + }) + + it('defaults off and follows the validated user layer', async () => { + const ctx = new Context() + await ctx.plugin(MemorySettings) + await ctx.plugin(SubagentModelSelectionConfig) + + expect(ctx.subagentModelSelection.currentEnabled()).toBe(false) + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: true }) + expect(ctx.subagentModelSelection.currentEnabled()).toBe(true) + await ctx.fiber.dispose() + }) + + it('samples each new root session without changing existing Agents', async () => { + const ctx = await boot() + const disabled = await createAgent(ctx, 'disabled') + expect(selectable(ctx, disabled)).toBe(false) + expect(hasSubagentModelSelection(disabled.session)).toBe(false) + + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: true }) + const enabled = await createAgent(ctx, 'enabled') + expect(hasSubagentModelSelection(enabled.session)).toBe(true) + expect(selectable(ctx, enabled)).toBe(true) + expect(selectable(ctx, disabled)).toBe(false) + + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: false }) + const disabledAgain = await createAgent(ctx, 'disabled-again') + expect(selectable(ctx, disabledAgain)).toBe(false) + expect(selectable(ctx, enabled)).toBe(true) + await ctx.fiber.dispose() + }) + + it('installs per-Agent definitions for a shared preset scope', async () => { + const ctx = await boot() + const preset = createScope(ctx, { preset: 'standard' }) + const other = createScope(ctx, { preset: 'minimal' }) + await preset.ctx.plugin(tool, { + provider: 'spawn', + modelSelectionSettings: true, + backgroundMode: 'continuable', + }) + + let enabledBinding: ReturnType | undefined + const createComposed = async (id: string) => ctx.agents.create({ + sessionId: SessionId(id), + setup: (agentCtx) => { + const binding = bindScopeParent(scopeOf(agentCtx)!, scopeOf(preset.ctx)!) + if (id === 'preset-enabled') enabledBinding = binding + }, + }) + + const disabled = await createComposed('preset-disabled') + expect(selectable(ctx, disabled.agent)).toBe(false) + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: true }) + const enabled = await createComposed('preset-enabled') + expect(selectable(ctx, enabled.agent)).toBe(true) + expect(selectable(ctx, disabled.agent)).toBe(false) + + enabledBinding!.rebind(scopeOf(other.ctx)!) + ctx.emit(scopeTarget({}, scopeOf(preset.ctx)), 'tools/change') + await vi.waitFor(() => { expect(selectable(ctx, enabled.agent)).toBe(false) }) + enabledBinding!.rebind(scopeOf(preset.ctx)!) + ctx.emit(scopeTarget({}, scopeOf(preset.ctx)), 'tools/change') + await vi.waitFor(() => { expect(selectable(ctx, enabled.agent)).toBe(true) }) + + await enabled.dispose() + ctx.emit(scopeTarget({}, scopeOf(preset.ctx)), 'tools/change') + await disabled.dispose() + await ctx.fiber.dispose() + }) + + it('inherits the parent decision and preserves seeded decisions across composition', async () => { + const ctx = await boot() + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: true }) + const parent = await createAgent(ctx, 'parent') + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: false }) + const child = await createAgent(ctx, 'child', { + meta: { parentSession: parent.id, origin: 'subagent' }, + }) + expect(selectable(ctx, child)).toBe(true) + expect(hasSubagentModelSelection(child.session)).toBe(true) + + const enabledSeed = Session.create(SessionId('enabled-seed')) + enabledSeed.append('subagent/model-selection-enabled', {}) + const resumedEnabled = await createAgent(ctx, 'resumed-enabled', { seed: enabledSeed.events }) + expect(selectable(ctx, resumedEnabled)).toBe(true) + + const oldSeed = Session.create(SessionId('old-seed'), []) + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: true }) + const resumedDisabled = await createAgent(ctx, 'resumed-disabled', { seed: oldSeed.events }) + expect(selectable(ctx, resumedDisabled)).toBe(false) + expect(hasSubagentModelSelection(resumedDisabled.session)).toBe(false) + await ctx.fiber.dispose() + }) + + it('rejects ambiguous static and settings-controlled configuration', async () => { + const ctx = new Context() + await mountAgentLoopTestDependencies(ctx) + await ctx.plugin(SubagentRuntime) + expect(() => { + tool.apply(ctx, { + provider: 'missing', + enableModelSelection: true, + modelSelectionSettings: true, + }) + }).toThrow('mutually exclusive') + await ctx.fiber.dispose() + }) + + it('requires both the Host setting owner and a composition scope', async () => { + const withoutSettings = new Context() + await mountAgentLoopTestDependencies(withoutSettings) + await withoutSettings.plugin(SubagentRuntime) + expect(() => { + tool.apply(withoutSettings, { + provider: 'missing', + modelSelectionSettings: true, + maxDepth: 'provider-managed', + }) + }).toThrow('requires @deepseek-ai/dsh-tool-subagent/model-selection-settings') + await withoutSettings.fiber.dispose() + + const withoutAgent = await boot() + expect(() => { + tool.apply(withoutAgent, { + provider: 'spawn', + modelSelectionSettings: true, + backgroundMode: 'continuable', + }) + }).toThrow('requires an Agent or preset scope') + await withoutAgent.fiber.dispose() + }) + + it('checks the durable decision against the published tool definitions', async () => { + const ctx = await boot() + await ctx.plugin(InvariantRegistry, { enabled: true }) + await ctx.plugin(ToolInvariant) + const disabled = await createAgent(ctx, 'invariant-disabled') + const next = () => Promise.resolve({ kind: 'enter' as const, messages: [] }) + const payload = { + agent: disabled, + messages: [], + turn: 1, + step: 1, + signal: new AbortController().signal, + } + await expect(ctx.waterfall(ctx as never, 'agent/pre-step', payload, next)).resolves.toEqual({ + kind: 'enter', messages: [], + }) + + disabled.session.append('subagent/model-selection-enabled', {}) + await expect(ctx.waterfall(ctx as never, 'agent/pre-step', payload, next)) + .rejects.toThrow('must expose route fields and list_subagent_models') + + await ctx.settings.update(SUBAGENT_MODEL_SELECTION_SETTINGS_NAMESPACE, { enabled: true }) + const enabled = await createAgent(ctx, 'invariant-enabled') + await expect(ctx.waterfall(ctx as never, 'agent/pre-step', { ...payload, agent: enabled }, next)) + .resolves.toEqual({ kind: 'enter', messages: [] }) + await ctx.fiber.dispose() + }) +}) diff --git a/packages/subagent/tool-subagent/tests/model-selection.spec.ts b/packages/subagent/tool-subagent/tests/model-selection.spec.ts new file mode 100644 index 0000000000..008ba66da4 --- /dev/null +++ b/packages/subagent/tool-subagent/tests/model-selection.spec.ts @@ -0,0 +1,365 @@ +import { describe, expect, it, vi } from 'vitest' +import { Context } from '@deepseek-ai/cordis' +import { ReasoningEffortId } from '@deepseek-ai/dsh-llm' +import SystemPrompt from '@deepseek-ai/dsh-system-prompt' +import ToolRuntime from '@deepseek-ai/dsh-tools' +import type { Agent } from '@deepseek-ai/dsh-agent' +import SubagentRuntime from '@deepseek-ai/dsh-subagent' +import type { SubagentStartRequest } from '@deepseek-ai/dsh-subagent' +import { Session, SessionId } from '@deepseek-ai/dsh-session' +import { MockAdapter } from '../../../core/agent-loop/tests/mock-adapter.ts' +import * as mock from './scripted-provider.ts' +import * as tool from '../src/index.ts' +import { callSubagent, setup, text } from './harness.ts' + +const REASONING = { + efforts: [ + { id: ReasoningEffortId('low'), name: 'Low' }, + { id: ReasoningEffortId('high'), name: 'High' }, + ], + defaultEffort: ReasoningEffortId('high'), +} as const + +function parentWithRoute( + options: Agent['options'] = { + provider: 'alpha', + model: 'parent-model', + reasoningEffort: ReasoningEffortId('high'), + }, +): Agent { + const id = SessionId('parent-with-route') + return { id, options, session: Session.create(id) } as unknown as Agent +} + +describe('dsh-tool-subagent model selection', () => { + it('exposes static route fields and discovery when selection is enabled', async () => { + const ctx = await setup({ provider: 'mock', enableModelSelection: true }) + const schema = ctx.tools.schemas().find(entry => entry.name === 'subagent')! + const props = (schema.parameters as { properties?: Record }).properties ?? {} + expect(Object.keys(props).sort()).toEqual([ + 'description', + 'model', + 'prompt', + 'provider', + 'reasoning_effort', + 'run_in_background', + ]) + expect(schema.description).toContain('list_subagent_models') + expect(ctx.tools.get('list_subagent_models')).toBeDefined() + expect(schema.description).not.toContain('alpha') + + const registration = ctx.llm.registerAdapter(['alpha'], new MockAdapter([])) + const definition = ctx.tools.get('subagent') + registration.replace(['beta']) + expect(ctx.tools.get('subagent')).toBe(definition) + expect(definition?.description).not.toContain('beta') + }) + + it('hides and rejects route fields when selection is disabled', async () => { + const ctx = await setup({ provider: 'mock' }) + const schema = ctx.tools.schemas().find(entry => entry.name === 'subagent')! + const props = (schema.parameters as { properties?: Record }).properties ?? {} + expect(Object.keys(props).sort()).toEqual(['description', 'prompt', 'run_in_background']) + expect(schema.description).not.toContain('list_subagent_models') + expect(ctx.tools.get('list_subagent_models')).toBeUndefined() + + const result = await callSubagent(ctx, { + description: 'forced route', + prompt: 'do it', + provider: 'alpha', + model: 'fast-model', + }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('child model selection is disabled for this tool instance') + }) + + it('rejects enabled model selection when the provider cannot apply Agent options', async () => { + await expect(setup( + { provider: 'mock', enableModelSelection: true, maxDepth: 'provider-managed' }, + { capabilities: { agentOptions: false } }, + )).rejects.toThrow('provider "mock" does not support child model selection') + }) + + it('selects an unlisted complete route and clears a configured effort when the route changes', async () => { + const requests: SubagentStartRequest[] = [] + const ctx = await setup({ + provider: 'mock', + enableModelSelection: true, + agentOptions: { + provider: 'alpha', + model: 'configured-model', + reasoningEffort: ReasoningEffortId('high'), + maxTokens: 321, + }, + }, { onStart: (request) => { requests.push(request) } }) + ctx.llm.registerAdapter(['alpha'], new MockAdapter([], REASONING)) + + const selected = await callSubagent(ctx, { + description: 'route work', + prompt: 'do it', + provider: 'alpha', + model: 'unlisted-model', + }) + expect(selected.isError).toBe(false) + expect(requests[0]?.agentOptions).toEqual({ + provider: 'alpha', + model: 'unlisted-model', + maxTokens: 321, + }) + + const effort = await callSubagent(ctx, { + description: 'same route effort', + prompt: 'do it', + provider: 'alpha', + model: 'configured-model', + reasoning_effort: 'low', + }) + expect(effort.isError).toBe(false) + expect(requests[1]?.agentOptions).toEqual({ + provider: 'alpha', + model: 'configured-model', + reasoningEffort: 'low', + maxTokens: 321, + }) + }) + + it('accepts an effort-only override for the effective configured or parent route', async () => { + const requests: SubagentStartRequest[] = [] + const ctx = await setup({ + provider: 'mock', + enableModelSelection: true, + agentOptions: { provider: 'alpha' }, + }, { onStart: (request) => { requests.push(request) } }) + ctx.llm.registerAdapter(['alpha'], new MockAdapter([], REASONING)) + + const result = await callSubagent(ctx, { + description: 'effort work', + prompt: 'do it', + reasoning_effort: 'low', + }, { agent: parentWithRoute() }) + expect(result.isError).toBe(false) + expect(requests[0]?.agentOptions).toEqual({ provider: 'alpha', reasoningEffort: 'low' }) + + const inherited = await setup({ provider: 'mock', enableModelSelection: true }) + inherited.llm.registerAdapter(['alpha'], new MockAdapter([], REASONING)) + const inheritedResult = await callSubagent(inherited, { + description: 'parent effort work', + prompt: 'do it', + reasoning_effort: 'low', + }, { agent: parentWithRoute() }) + expect(inheritedResult.isError).toBe(false) + }) + + it('inherits a parent effort only when an explicit route stays unchanged', async () => { + const ctx = await setup({ provider: 'mock', enableModelSelection: true }) + ctx.llm.registerAdapter(['alpha'], new MockAdapter([], REASONING)) + const result = await callSubagent(ctx, { + description: 'same route work', + prompt: 'do it', + provider: 'alpha', + model: 'parent-model', + }, { agent: parentWithRoute() }) + expect(result.isError).toBe(false) + }) + + it('compares explicit routes with the latest logged parent selection', async () => { + const requests: SubagentStartRequest[] = [] + const ctx = await setup({ + provider: 'mock', + enableModelSelection: true, + agentOptions: { reasoningEffort: ReasoningEffortId('high') }, + }, { onStart: (request) => { requests.push(request) } }) + ctx.llm.registerAdapter(['current-provider'], new MockAdapter([], REASONING)) + const parent = parentWithRoute({ provider: 'created-provider', model: 'created-model' }) + parent.session.append('request/header', { + header: { config: { provider: 'current-provider', model: 'current-model' } }, + reason: 'initial', + }) + + const result = await callSubagent(ctx, { + description: 'same current route', + prompt: 'do it', + provider: 'current-provider', + model: 'current-model', + }, { agent: parent }) + + expect(result.isError).toBe(false) + expect(requests[0]?.agentOptions).toEqual({ + provider: 'current-provider', + model: 'current-model', + reasoningEffort: 'high', + }) + }) + + it('rejects an effort without any effective route', async () => { + const ctx = await setup({ provider: 'mock', enableModelSelection: true }) + const result = await callSubagent(ctx, { + description: 'missing route', + prompt: 'do it', + reasoning_effort: 'low', + }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('without an effective provider and model') + }) + + it.each([ + { provider: 'alpha' }, + { model: 'fast-model' }, + ])('rejects a partial model-facing route before child creation', async (route) => { + let starts = 0 + const ctx = await setup({ provider: 'mock', enableModelSelection: true }, { onStart: () => { starts += 1 } }) + const result = await callSubagent(ctx, { description: 'partial route', prompt: 'do it', ...route }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('`provider` and `model` must be supplied together') + expect(starts).toBe(0) + }) + + it.each([ + { provider: '', model: 'fast-model', expected: '`provider` must be non-empty' }, + { provider: 'alpha', model: '', expected: '`model` must be non-empty' }, + { reasoning_effort: '', expected: '`reasoning_effort` must be non-empty' }, + ])('rejects empty model-facing values', async ({ expected, ...selection }) => { + const ctx = await setup({ provider: 'mock', enableModelSelection: true }) + const result = await callSubagent(ctx, { description: 'empty route', prompt: 'do it', ...selection }) + expect(result.isError).toBe(true) + expect(text(result)).toContain(expected) + }) + + it('uses the LLM runtime for provider and reasoning-effort validation before child creation', async () => { + let starts = 0 + const ctx = await setup({ provider: 'mock', enableModelSelection: true }, { onStart: () => { starts += 1 } }) + ctx.llm.registerAdapter(['alpha'], new MockAdapter([], REASONING)) + + const unsupported = await callSubagent(ctx, { + description: 'bad effort', + prompt: 'do it', + provider: 'alpha', + model: 'fast-model', + reasoning_effort: 'max', + }) + expect(unsupported.isError).toBe(true) + expect(text(unsupported)).toContain('does not support reasoning effort "max"') + + const missing = await callSubagent(ctx, { + description: 'bad provider', + prompt: 'do it', + provider: 'missing', + model: 'fast-model', + }) + expect(missing.isError).toBe(true) + expect(text(missing)).toContain('no adapter registered for provider "missing"') + expect(starts).toBe(0) + }) + + it('validates a configured effort before child creation', async () => { + let starts = 0 + const ctx = await setup({ + provider: 'mock', + agentOptions: { + provider: 'alpha', + model: 'parent-model', + reasoningEffort: ReasoningEffortId('high'), + }, + }, { onStart: () => { starts += 1 } }) + ctx.llm.registerAdapter(['alpha'], new MockAdapter([], { + efforts: [{ id: ReasoningEffortId('low'), name: 'Low' }], + defaultEffort: ReasoningEffortId('low'), + })) + + const result = await callSubagent( + ctx, + { description: 'same route', prompt: 'do it' }, + { agent: parentWithRoute() }, + ) + expect(result.isError).toBe(true) + expect(text(result)).toContain('does not support reasoning effort "high"') + expect(starts).toBe(0) + }) + + it('validates a configured route before child creation', async () => { + let starts = 0 + const ctx = await setup({ + provider: 'mock', + agentOptions: { provider: 'missing', model: 'configured-model' }, + }, { onStart: () => { starts += 1 } }) + + const result = await callSubagent( + ctx, + { description: 'configured route', prompt: 'do it' }, + { agent: parentWithRoute() }, + ) + + expect(result.isError).toBe(true) + expect(text(result)).toContain('no adapter registered for provider "missing"') + expect(starts).toBe(0) + }) + + it('rejects selected routes or configured efforts when the LLM service is absent', async () => { + const ctx = new Context() + await ctx.plugin(SystemPrompt) + await ctx.plugin(ToolRuntime) + await ctx.plugin(SubagentRuntime) + await mock.mountScriptedProvider(ctx, { name: 'mock' }) + await ctx.plugin(tool, { + provider: 'mock', + enableModelSelection: true, + agentOptions: { + provider: 'alpha', + model: 'fast-model', + reasoningEffort: ReasoningEffortId('high'), + }, + }) + + const configured = await callSubagent(ctx, { description: 'configured effort', prompt: 'do it' }) + expect(configured.isError).toBe(true) + expect(text(configured)).toContain('`llm` service is unavailable') + + const selected = await callSubagent(ctx, { + description: 'selected route', + prompt: 'do it', + provider: 'alpha', + model: 'other-model', + }) + expect(selected.isError).toBe(true) + expect(text(selected)).toContain('`llm` service is unavailable') + }) + + it('keeps pure inherited routing usable without an LLM service lookup', async () => { + let starts = 0 + const ctx = new Context() + await ctx.plugin(SystemPrompt) + await ctx.plugin(ToolRuntime) + await ctx.plugin(SubagentRuntime) + await mock.mountScriptedProvider(ctx, { name: 'mock', onStart: () => { starts += 1 } }) + await ctx.plugin(tool, { provider: 'mock' }) + + const result = await callSubagent(ctx, { description: 'inherit route', prompt: 'do it' }) + expect(result.isError).toBe(false) + expect(starts).toBe(1) + }) + + it('warns that changing a fork route can lose inherited-prefix reuse', async () => { + const ctx = await setup({ provider: 'mock', enableModelSelection: true }, { inheritsParentContext: true }) + const schema = ctx.tools.schemas().find(entry => entry.name === 'subagent')! + expect(schema.description).toContain('inherits this conversation') + expect(schema.description).toContain('can prevent provider-side reuse of the inherited conversation prefix') + }) + + it('propagates an exact-route resolver failure before child creation', async () => { + let starts = 0 + const ctx = await setup({ provider: 'mock', enableModelSelection: true }, { onStart: () => { starts += 1 } }) + const adapter = new MockAdapter([]) + vi.spyOn(adapter, 'resolveModel').mockRejectedValue(new Error('selected route unavailable')) + ctx.llm.registerAdapter(['alpha'], adapter) + + const result = await callSubagent(ctx, { + description: 'route work', + prompt: 'do it', + provider: 'alpha', + model: 'fast-model', + }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('selected route unavailable') + expect(starts).toBe(0) + }) +}) diff --git a/packages/subagent/tool-subagent/tests/scripted-provider.ts b/packages/subagent/tool-subagent/tests/scripted-provider.ts index c0da403cd4..a946c6f6fe 100644 --- a/packages/subagent/tool-subagent/tests/scripted-provider.ts +++ b/packages/subagent/tool-subagent/tests/scripted-provider.ts @@ -13,6 +13,7 @@ import type { } from '@deepseek-ai/dsh-subagent' const DEFAULT_CAPABILITIES: SubagentCapabilities = { + agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, diff --git a/packages/subagent/tool-subagent/tests/tool-subagent.spec.ts b/packages/subagent/tool-subagent/tests/tool-subagent.spec.ts index 1ee5e40228..334b721461 100644 --- a/packages/subagent/tool-subagent/tests/tool-subagent.spec.ts +++ b/packages/subagent/tool-subagent/tests/tool-subagent.spec.ts @@ -4,7 +4,7 @@ import { tmpdir } from 'node:os' import path from 'node:path' import { Context } from '@deepseek-ai/cordis' import Loader from '@deepseek-ai/cordis-plugin-loader' -import { CallId } from '@deepseek-ai/dsh-llm' +import LlmRuntime, { CallId } from '@deepseek-ai/dsh-llm' import SystemPrompt from '@deepseek-ai/dsh-system-prompt' import ToolRuntime, { TOOL_ABORTED_BEFORE_DISPATCH } from '@deepseek-ai/dsh-tools' import { assembleContextFor, type Agent } from '@deepseek-ai/dsh-agent' @@ -20,9 +20,8 @@ import * as ToolTasks from '@deepseek-ai/dsh-tool-jobs' import { MockAdapter, textResponse } from '../../../core/agent-loop/tests/mock-adapter.ts' import * as mock from './scripted-provider.ts' import * as tool from '../src/index.ts' -import { SessionId } from '@deepseek-ai/dsh-session' - -const testToolSignal = new AbortController().signal +import { Session, SessionId } from '@deepseek-ai/dsh-session' +import { callSubagent, fakeAgent, setup, testToolSignal, text } from './harness.ts' /** * Drives the REAL plugin body: mounts `dsh-tool-subagent` on a real @@ -32,40 +31,6 @@ const testToolSignal = new AbortController().signal * shipping code path. */ -/** A minimal parent Agent passed through to the provider request. */ -function fakeAgent(id = 'parent-1'): Agent { - return { id: SessionId(id) } as unknown as Agent -} - -async function setup(toolConfig: tool.Config, mockConfig: Partial = {}) { - const ctx = new Context() - await ctx.plugin(SystemPrompt) - await ctx.plugin(ToolRuntime) - await ctx.plugin(SubagentRuntime) - await mock.mountScriptedProvider(ctx, { name: 'mock', ...mockConfig }) - await ctx.plugin(tool, toolConfig) - return ctx -} - -let callCounter = 0 -function callSubagent(ctx: Context, args: unknown, over: { agent?: Agent | undefined; signal?: AbortSignal } = {}) { - // Distinguish "no override" (use a default agent) from an explicit - // `{ agent: undefined }` (test the no-agent path). Under - // exactOptionalPropertyTypes the key is omitted rather than set to undefined. - const agent = 'agent' in over ? over.agent : fakeAgent() - return ctx.tools.execute({ - signal: testToolSignal, - callId: CallId(`call-${++callCounter}`), - name: 'subagent', - arguments: args, - ...agent ? { agent } : {}, - ...over.signal ? { signal: over.signal } : {}, - }) -} - -function text(result: { content: { type: string; text?: string }[] }): string { - return result.content.filter(b => b.type === 'text').map(b => b.text).join('') -} describe('dsh-tool-subagent', () => { it('rejects continuable background policy when the provider cannot prepare continuable children', async () => { @@ -83,6 +48,13 @@ describe('dsh-tool-subagent', () => { ) }) + it('rejects configured child agent options at mount when the provider cannot apply them', async () => { + await expect(setup( + { provider: 'mock', maxDepth: 'provider-managed', agentOptions: { model: 'configured-model' } }, + { capabilities: { agentOptions: false } }, + )).rejects.toThrow('does not support child agentOptions') + }) + it('registers a `subagent` tool that delegates to the configured provider and returns its output', async () => { const ctx = await setup({ provider: 'mock' }, { reply: 'child says hi' }) const result = await callSubagent(ctx, { @@ -100,20 +72,14 @@ describe('dsh-tool-subagent', () => { expect(text(result)).toBe('child says hi') }) - it('exposes description + prompt + run_in_background to the model (no provider/type parameter)', async () => { - const ctx = await setup({ provider: 'mock' }) - const schema = ctx.tools.schemas().find(s => s.name === 'subagent') - expect(schema).toBeDefined() - const props = (schema!.parameters as { properties?: Record }).properties ?? {} - expect(Object.keys(props).sort()).toEqual(['description', 'prompt', 'run_in_background']) - expect(schema!.description).toContain('job_output') - }) - it('omits run_in_background entirely when the instance disables it (schema and capability never disagree)', async () => { const ctx = await setup({ provider: 'mock', enableRunInBackground: false }) const schema = ctx.tools.schemas().find(s => s.name === 'subagent') const props = (schema!.parameters as { properties?: Record }).properties ?? {} - expect(Object.keys(props).sort()).toEqual(['description', 'prompt']) + expect(Object.keys(props).sort()).toEqual([ + 'description', + 'prompt', + ]) expect(schema!.description).not.toContain('job_output') }) @@ -121,7 +87,13 @@ describe('dsh-tool-subagent', () => { // Schema omission is advertising, not enforcement: the arg validator // allows undeclared keys, so the opt-out must also hold in execute(). const ctx = await setup({ provider: 'mock', enableRunInBackground: false }) - const parent = { id: SessionId('sess-off'), inject: () => {}, options: {}, session: { header: { version: 0, id: 'sess-off', createdAt: 0 } } } as unknown as Agent + const parentId = SessionId('sess-off') + const parent = { + id: parentId, + inject: () => {}, + options: {}, + session: Session.create(parentId), + } as unknown as Agent const forced = await callSubagent(ctx, { description: 'd', prompt: 'p', run_in_background: true }, { agent: parent }) expect(forced.isError).toBe(true) @@ -232,7 +204,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'weird', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => ({ id: SessionId('weird-child'), @@ -253,12 +225,13 @@ describe('dsh-tool-subagent', () => { // the request lets us assert the agentOptions reached it. let seen: { agentOptions?: { model?: string } } | undefined const ctx = new Context() + await ctx.plugin(LlmRuntime) await ctx.plugin(SystemPrompt) await ctx.plugin(ToolRuntime) await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'capture', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: true, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async (request) => { seen = request @@ -270,10 +243,15 @@ describe('dsh-tool-subagent', () => { } }, }) - await ctx.plugin(tool, { provider: 'capture', agentOptions: { model: 'child-model' }, maxDepth: 'provider-managed' }) + ctx.llm.registerAdapter(['alpha'], new MockAdapter([])) + await ctx.plugin(tool, { + provider: 'capture', + agentOptions: { provider: 'alpha', model: 'child-model' }, + maxDepth: 'provider-managed', + }) await callSubagent(ctx, { description: 'd', prompt: 'p' }) - expect(seen?.agentOptions).toEqual({ model: 'child-model' }) + expect(seen?.agentOptions).toEqual({ provider: 'alpha', model: 'child-model' }) }) it('defaults toolName and omits agentOptions when apply() is called directly (schema bypass)', async () => { @@ -288,7 +266,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'bare', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async (request) => { seen = request @@ -378,7 +356,7 @@ describe('dsh-tool-subagent', () => { // the provider survives. ctx.subagents.registerProvider({ name: 'continuable', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => { throw new Error('lifecycle test does not start a child') }, prepareContinuable: async () => ({}), @@ -429,10 +407,14 @@ describe('dsh-tool-subagent', () => { }) it('derives inherited-context wording from a seeded-conversation provider', async () => { - const ctx = await setup({ provider: 'mock', toolName: 'subagent' }, { inheritsParentContext: true }) + const ctx = await setup({ + provider: 'mock', + toolName: 'subagent', + }, { inheritsParentContext: true }) const schema = ctx.tools.schemas().find(s => s.name === 'subagent')! expect(schema.description).toContain('inherits this conversation') expect(schema.description).not.toContain('does not see this conversation') + expect(schema.description).not.toContain('can prevent provider-side reuse of the inherited conversation prefix') const props = (schema.parameters as { properties: Record }).properties expect(props['prompt']!.description).toContain('completed turns') }) @@ -447,7 +429,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'spy', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => ({ id: SessionId('spy-child'), @@ -470,7 +452,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'spy', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => ({ id: SessionId('spy-child'), @@ -494,7 +476,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'spy', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => ({ id: SessionId('spy-child'), @@ -522,7 +504,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'spy', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => ({ id: SessionId('spy-child'), @@ -549,7 +531,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'spy', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async (request) => { if (request.signal.aborted) throw new Error('start aborted') @@ -588,7 +570,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'spy', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async (request) => { if (request.signal.aborted) sawAborted() @@ -652,7 +634,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'capture2', - capabilities: { outputSchema: false, depthLimit: true, toolFilter: true, persona: true }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: true, toolFilter: true, persona: true }, inheritsParentContext: false, start: async (request) => { seen = request @@ -709,7 +691,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'capture3', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: true, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: true, persona: false }, inheritsParentContext: false, start: async (request) => { seen = request @@ -739,7 +721,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'capture4', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async (request) => { seen = request @@ -764,7 +746,7 @@ describe('dsh-tool-subagent', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'p', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: true, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: true, persona: false }, inheritsParentContext: false, start: () => { throw new Error('unreachable') }, }) @@ -783,7 +765,7 @@ describe('dsh-tool-subagent background mode', () => { ctx: scopeFiber.ctx, inject, options: {}, - session: { id, header: { version: 0, id, createdAt: 0 } }, + session: Session.create(id), } as unknown as Agent ctx.agents.register(agent) return agent @@ -803,7 +785,7 @@ describe('dsh-tool-subagent background mode', () => { let prepareCalls = 0 ctx.subagents.registerProvider({ name: 'resumable', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async request => ({ id: SessionId('one-shot-child'), @@ -839,7 +821,7 @@ describe('dsh-tool-subagent background mode', () => { }) it('returns a job id immediately and the answer is collected through job_output', async () => { - const ctx = await backgroundSetup({ provider: 'mock', agentOptions: { model: 'child-model' } }, { reply: 'background answer' }) + const ctx = await backgroundSetup({ provider: 'mock' }, { reply: 'background answer' }) const parent = ownerAgent(ctx, 'sess-parent') const start = await callSubagent(ctx, { description: 'deep research', prompt: 'dig in', run_in_background: true }, { agent: parent }) @@ -919,12 +901,41 @@ describe('dsh-tool-subagent background mode', () => { expect(text(result)).toBe('Error: tool call aborted before dispatch') }) + it('skips background startup when cancellation wins asynchronous route preflight', async () => { + const ctx = await backgroundSetup({ provider: 'mock', enableModelSelection: true }) + const parent = ownerAgent(ctx, 'sess-parent') + const adapter = new MockAdapter([]) + let releasePreflight!: () => void + const preflightGate = new Promise((resolve) => { releasePreflight = resolve }) + const resolveModel = vi.spyOn(adapter, 'resolveModel').mockImplementation(async (provider, model) => { + await preflightGate + return { provider, id: model, name: model } + }) + ctx.llm.registerAdapter(['alpha'], adapter) + const controller = new AbortController() + + const resultPromise = callSubagent(ctx, { + description: 'cancelled selection', + prompt: 'do it', + provider: 'alpha', + model: 'selected-model', + run_in_background: true, + }, { agent: parent, signal: controller.signal }) + await vi.waitFor(() => { expect(resolveModel).toHaveBeenCalledOnce() }) + controller.abort() + releasePreflight() + const result = await resultPromise + + expect(result.isError).toBe(true) + expect(ctx.jobs.list(parent)).toEqual([]) + }) + it('settles an asynchronous provider-start failure as a failed task', async () => { const ctx = await backgroundSetup({ provider: 'mock' }) const parent = ownerAgent(ctx, 'sess-parent') ctx.subagents.registerProvider({ name: 'broken-start', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => { throw new Error('setup failed') }, }) @@ -953,7 +964,7 @@ describe('dsh-tool-subagent background mode', () => { const parent = ownerAgent(ctx, 'sess-parent') ctx.subagents.registerProvider({ name: 'pending-start', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: request => new Promise((_resolve, reject) => { request.signal.addEventListener('abort', () => { reject(new Error('startup aborted')) }, { once: true }) @@ -990,7 +1001,7 @@ describe('dsh-tool-subagent background mode', () => { const parent = ownerAgent(ctx, 'sess-parent') ctx.subagents.registerProvider({ name: 'broken-start-rollback', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: request => new Promise((_resolve, reject) => { request.signal.addEventListener('abort', () => { @@ -1035,7 +1046,7 @@ describe('dsh-tool-subagent background mode', () => { let starts = 0 ctx.subagents.registerProvider({ name: 'hanging', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async (request) => { let settle!: (value: { output: { type: 'text'; text: string }[]; stopReason: 'aborted' }) => void @@ -1182,7 +1193,7 @@ describe('dsh-tool-subagent continuable background mode', () => { let survivingChildId: ReturnType | undefined ctx.subagents.registerProvider({ name: 'gated', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, + capabilities: { agentOptions: false, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, inheritsParentContext: false, start: async () => { throw new Error('continuable policy must not start a one-shot child') }, prepareContinuable: async (request) => { @@ -1250,14 +1261,14 @@ describe('background preflight failure (no orphaned child, by construction)', () ctx: scopeFiber.ctx, inject: () => {}, options: {}, - session: { id, header: { version: 0, id, createdAt: 0 } }, + session: Session.create(id), } as unknown as Agent ctx.agents.register(parent) let starts = 0 ctx.subagents.registerProvider({ name: 'probe', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => { starts += 1 @@ -1295,7 +1306,7 @@ describe('depth budget configuration', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'capture', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, + capabilities: { agentOptions: false, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, inheritsParentContext: false, start: async (request) => { requests.push(request) @@ -1333,7 +1344,7 @@ describe('depth budget configuration', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'no-depth', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async () => { throw new Error('unreachable') }, }) @@ -1349,7 +1360,7 @@ describe('depth budget configuration', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'external', - capabilities: { outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: false, outputSchema: false, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, start: async (request) => { requests.push(request) diff --git a/packages/subagent/tool-subagent/tsconfig.json b/packages/subagent/tool-subagent/tsconfig.json index 8ee77dfb3b..57ee1dc9d6 100644 --- a/packages/subagent/tool-subagent/tsconfig.json +++ b/packages/subagent/tool-subagent/tsconfig.json @@ -20,6 +20,15 @@ { "path": "../../core/agent" }, + { + "path": "../../core/session" + }, + { + "path": "../../core/scope" + }, + { + "path": "../../settings/settings" + }, { "path": "../../llm/llm" }, diff --git a/packages/subagent/tool-subagent/tsdown.config.ts b/packages/subagent/tool-subagent/tsdown.config.ts new file mode 100644 index 0000000000..d91febcc25 --- /dev/null +++ b/packages/subagent/tool-subagent/tsdown.config.ts @@ -0,0 +1,19 @@ +import { defineConfig } from 'tsdown' + +const entry = (path: string) => ({ + entry: [path], + outDir: 'lib', + format: ['esm'] as const, + platform: 'node' as const, + target: 'es2024' as const, + fixedExtension: false, + dts: false, + clean: false, +}) + +/** Build self-contained Loader entries so the package needs no private chunks. */ +export default defineConfig([ + entry('lib/types/index.js'), + entry('lib/types/model-selection-settings.js'), + entry('lib/types/invariant.js'), +]) diff --git a/packages/workflow/tool-ralph/tests/tool-ralph.spec.ts b/packages/workflow/tool-ralph/tests/tool-ralph.spec.ts index 7c6f02e36c..0ccbbd16c6 100644 --- a/packages/workflow/tool-ralph/tests/tool-ralph.spec.ts +++ b/packages/workflow/tool-ralph/tests/tool-ralph.spec.ts @@ -56,6 +56,7 @@ class StubProvider implements SubagentProvider { constructor(options?: { outputSchema?: boolean; inheritsParentContext?: boolean }) { this.capabilities = { + agentOptions: true, outputSchema: options?.outputSchema ?? true, depthLimit: true, toolFilter: true, diff --git a/packages/workflow/tool-workflow/tests/tool-workflow.spec.ts b/packages/workflow/tool-workflow/tests/tool-workflow.spec.ts index 938f9a5502..f89083faba 100644 --- a/packages/workflow/tool-workflow/tests/tool-workflow.spec.ts +++ b/packages/workflow/tool-workflow/tests/tool-workflow.spec.ts @@ -423,7 +423,7 @@ describe('dsh-tool-workflow', () => { await ctx.plugin(SubagentRuntime) ctx.subagents.registerProvider({ name: 'spawn', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, inheritsParentContext: false, start: () => Promise.reject(new Error('the parked-script fixture must not start a child')), }) diff --git a/packages/workflow/workflow-worker-thread/tests/built-worker.e2e.ts b/packages/workflow/workflow-worker-thread/tests/built-worker.e2e.ts index 31998a22cf..cf7439e8eb 100644 --- a/packages/workflow/workflow-worker-thread/tests/built-worker.e2e.ts +++ b/packages/workflow/workflow-worker-thread/tests/built-worker.e2e.ts @@ -30,7 +30,7 @@ await ctx.plugin(SubagentRuntime) let selectedStarts = 0 ctx.subagents.registerProvider({ name: 'built-selected', - capabilities: { outputSchema: true, depthLimit: false, toolFilter: false, persona: false }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: false, toolFilter: false, persona: false }, inheritsParentContext: false, async start() { selectedStarts += 1 diff --git a/packages/workflow/workflow-worker-thread/tests/source-worker.compat.spec.ts b/packages/workflow/workflow-worker-thread/tests/source-worker.compat.spec.ts index 8d33e373c7..77b87b59f8 100644 --- a/packages/workflow/workflow-worker-thread/tests/source-worker.compat.spec.ts +++ b/packages/workflow/workflow-worker-thread/tests/source-worker.compat.spec.ts @@ -21,7 +21,7 @@ it('runs the default config through the source worker', async () => { const subagents = await ctx.plugin(SubagentRuntime) const provider: SubagentProvider = { name: 'spawn', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, inheritsParentContext: false, start: () => Promise.reject(new Error('source-worker compat script must not start a child')), } diff --git a/packages/workflow/workflow-worker-thread/tests/workflow-worker-thread.spec.ts b/packages/workflow/workflow-worker-thread/tests/workflow-worker-thread.spec.ts index 5130071362..935dbae3fb 100644 --- a/packages/workflow/workflow-worker-thread/tests/workflow-worker-thread.spec.ts +++ b/packages/workflow/workflow-worker-thread/tests/workflow-worker-thread.spec.ts @@ -55,7 +55,13 @@ interface ControlledRun { * the request signal fires, like the real in-process backends. */ class StubProvider implements SubagentProvider { - readonly capabilities: SubagentCapabilities = { outputSchema: true, depthLimit: true, toolFilter: true, persona: false } + readonly capabilities: SubagentCapabilities = { + agentOptions: true, + outputSchema: true, + depthLimit: true, + toolFilter: true, + persona: false, + } readonly inheritsParentContext = false readonly runs: ControlledRun[] = [] @@ -459,7 +465,7 @@ describe('dsh-workflow-worker-thread', () => { await ctx.plugin(SubagentRuntime) const provider: SubagentProvider = { name: 'rejecting', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, inheritsParentContext: false, start: async () => ({ id: SessionId('reject-child'), @@ -518,7 +524,7 @@ describe('dsh-workflow-worker-thread', () => { await ctx.plugin(SubagentRuntime) const provider: SubagentProvider = { name: 'bad-dispose', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, inheritsParentContext: false, start: async () => ({ id: SessionId('bad-dispose-child'), @@ -540,7 +546,7 @@ describe('dsh-workflow-worker-thread', () => { await ctx.plugin(SubagentRuntime) const provider: SubagentProvider = { name: 'coercion-trap-dispose', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, inheritsParentContext: false, start: async () => ({ id: SessionId('trap-child'), @@ -891,7 +897,7 @@ describe('dsh-workflow-worker-thread', () => { const aborted: string[] = [] const provider: SubagentProvider = { name: 'signal-only', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, inheritsParentContext: false, start: async (request) => { let settle!: (result: SubagentResult) => void @@ -1189,7 +1195,7 @@ describe('dsh-workflow-worker-thread', () => { const warn = vi.spyOn(ctx.logger, 'warn').mockImplementation(() => ctx.logger) const provider: SubagentProvider = { name: 'late-ready', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, inheritsParentContext: false, start: (request) => { requested.resolve(request) @@ -1250,7 +1256,7 @@ describe('dsh-workflow-worker-thread', () => { const signalAborts: unknown[] = [] const provider: SubagentProvider = { name: 'doomed', - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: false }, inheritsParentContext: false, start: async (request) => { request.signal.addEventListener('abort', () => { diff --git a/pnpm-lock.yaml b/pnpm-lock.yaml index a6b3e12d45..f6878a775b 100644 --- a/pnpm-lock.yaml +++ b/pnpm-lock.yaml @@ -1808,6 +1808,9 @@ importers: '@deepseek-ai/dsh-subprocess': specifier: workspace:^ version: link:../../subprocess/subprocess + '@deepseek-ai/dsh-tool-subagent': + specifier: workspace:^ + version: link:../../subagent/tool-subagent '@deepseek-ai/dsh-web-frontend': specifier: workspace:^ version: link:../../../apps/web @@ -8563,6 +8566,9 @@ importers: '@deepseek-ai/dsh-llm': specifier: workspace:^ version: link:../../llm/llm + '@deepseek-ai/dsh-scope': + specifier: workspace:^ + version: link:../../core/scope '@deepseek-ai/dsh-session': specifier: workspace:^ version: link:../../core/session @@ -8572,6 +8578,9 @@ importers: '@deepseek-ai/dsh-session-persistence-jsonl': specifier: workspace:^ version: link:../../session/session-persistence-jsonl + '@deepseek-ai/dsh-settings': + specifier: workspace:^ + version: link:../../settings/settings '@deepseek-ai/dsh-subagent': specifier: workspace:^ version: link:../subagent diff --git a/scripts/check-workspace-constraints.ts b/scripts/check-workspace-constraints.ts index 35e6eb1049..cd9620e0fa 100644 --- a/scripts/check-workspace-constraints.ts +++ b/scripts/check-workspace-constraints.ts @@ -153,6 +153,9 @@ const packageFileExtras: Readonly> = { '@deepseek-ai/dsh-code-runtime-python': ['py/**/*.py'], // The shipped preset compositions travel inside the roster package. '@deepseek-ai/dsh-agent-presets': ['presets'], + // The Web Host mounts the default-off settings owner independently of each + // Agent-scoped delegation-tool instance. + '@deepseek-ai/dsh-tool-subagent': ['lib/model-selection-settings.js'], // The argv-prefix runner entry ships beside the lib as its own bundle; // sandbox-local resolves it through the package's ./runner export. tsdown // also shares its generated FFI code through a hashed runtime chunk. diff --git a/scripts/gen-cordis-catalog.ts b/scripts/gen-cordis-catalog.ts index db52420840..05a4c8947c 100644 --- a/scripts/gen-cordis-catalog.ts +++ b/scripts/gen-cordis-catalog.ts @@ -100,6 +100,7 @@ export const SERVICE_PAGE: Record = { spillStore: 'spill.md', storage: 'storage.md', storageDomain: 'storage.md', + subagentModelSelection: 'subagent.md', subagents: 'subagent.md', subprocess: 'subprocess.md', systemPrompt: 'system-prompt.md', diff --git a/scripts/gen-doc-graphs.ts b/scripts/gen-doc-graphs.ts index c5452d8d4a..a47f94b943 100644 --- a/scripts/gen-doc-graphs.ts +++ b/scripts/gen-doc-graphs.ts @@ -205,6 +205,14 @@ const SERVICE_ROLES: ServiceRole[] = [ consumers: ['llm-deepseek', 'llm-pi-ai', 'apiproxy'], note: 'Plugins register namespace schemas and resolve layered values; providers store the raw document. The LLM adapters register their entry config as the composition base under the user section; the web gateway serves redacted layered descriptors and writes the user layer.', }, + { + key: 'subagentModelSelection', + pkg: 'tool-subagent', + title: 'Subagent model-selection preference', + mode: 'core', + consumers: ['tool-subagent'], + note: 'Owns the default-off settings namespace that Agent-scoped delegation tools sample when composing a new top-level Session.', + }, { key: 'credentials', pkg: 'credentials', diff --git a/scripts/gen-tool-catalog.ts b/scripts/gen-tool-catalog.ts index 19ab10c455..3d69da2cc3 100644 --- a/scripts/gen-tool-catalog.ts +++ b/scripts/gen-tool-catalog.ts @@ -9,6 +9,7 @@ import { globSync, readFileSync, writeFileSync } from 'node:fs' import { basename, resolve } from 'node:path' import { Context } from '@deepseek-ai/cordis' +import LlmRuntime from '@deepseek-ai/dsh-llm' import type { ToolSchema } from '@deepseek-ai/dsh-llm' import AgentRegistry from '@deepseek-ai/dsh-agent' import type { Agent } from '@deepseek-ai/dsh-agent' @@ -104,7 +105,7 @@ const OUT = 'docs/tool-catalog.md' function registerCatalogSubagentProvider(ctx: Context, name: string): void { const provider: SubagentProvider = { name, - capabilities: { outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, + capabilities: { agentOptions: true, outputSchema: true, depthLimit: true, toolFilter: true, persona: true }, inheritsParentContext: false, start: () => Promise.reject(new Error('tool-catalog provider cannot start a child')), // Declared so consumers configured for continuable background mode mount. @@ -455,17 +456,21 @@ const TOOL_PACKAGES: ToolPackage[] = [ { pkg: '@deepseek-ai/dsh-tool-subagent', dir: 'tool-subagent', - source: 'packages/subagent/tool-subagent/src/index.ts', - requires: ['ctx.tools', 'ctx.subagents', 'ctx.systemPrompt'], + source: { + list_subagent_models: 'packages/subagent/tool-subagent/src/list-models.ts', + subagent: 'packages/subagent/tool-subagent/src/index.ts', + }, + requires: ['ctx.tools', 'ctx.subagents', 'ctx.systemPrompt', 'ctx.llm for model discovery and selected-route validation'], writes: ['tool/call', 'tool/result', 'child session events through the chosen provider'], shippedNames: ['subagent', 'subagent_fork'], async mount(ctx) { await ctx.plugin(SubagentRuntime) + await ctx.plugin(LlmRuntime) registerCatalogSubagentProvider(ctx, 'mock') - await ctx.plugin(ToolSubagent, { provider: 'mock' }) + await ctx.plugin(ToolSubagent, { provider: 'mock', enableModelSelection: true }) }, note: - 'The registered tool name is the load-time `toolName` config (default `subagent`); the schema above is that default. The shipped compositions load this package once per subagent backend, so the model additionally sees `subagent_fork` bound to the fork backend. Each instance\'s description, `run_in_background` parameter, and system-prompt policy follow its own `backgroundMode` and `enableRunInBackground`, so the two shipped schemas are not identical: `subagent` is `continuable` and defaults omitted calls to background with automatic settlement delivery, while `subagent_fork` stays `one-shot` and defaults them to foreground — see `packages/bundle/base/cordis.patch.yml` and `examples/acp-agent/cordis.yml`.', + 'The registered delegation name is the load-time `toolName` config (default `subagent`); the schema above shows static model selection enabled for reference. Model selection defaults off. Web presets sample the default-off Models preference for each new top-level Session and preserve that decision for its child Sessions; `subagent_fork` remains fixed-route. Explicit compositions may instead use static `enableModelSelection`. Each instance independently controls model selection, discovery ownership, and background behavior through `enableModelSelection`, `modelSelectionSettings`, `backgroundMode`, and `enableRunInBackground`.', }, { pkg: '@deepseek-ai/dsh-tool-subagent-control', diff --git a/scripts/snapshots/python-sdk-single-exe/advanced/result.json b/scripts/snapshots/python-sdk-single-exe/advanced/result.json index 20aecae839..c9dc0735ea 100644 --- a/scripts/snapshots/python-sdk-single-exe/advanced/result.json +++ b/scripts/snapshots/python-sdk-single-exe/advanced/result.json @@ -2988,7 +2988,7 @@ "seq": 6, "time": 0, "data": { - "version": 2, + "version": 3, "mode": "one-shot", "provider": "spawn", "label": "Check direct child" @@ -3110,11 +3110,10 @@ "config": { "provider": "deepseek-official", "model": "smoke-model", - "maxTokens": 256000, - "reasoningEffort": "high" + "reasoningEffort": "high", + "maxTokens": 256000 }, "adapterDefaults": { - "reasoningEffort": true, "maxTokens": true }, "system": "{{system}}", @@ -3763,7 +3762,7 @@ "seq": 6, "time": 0, "data": { - "version": 2, + "version": 3, "mode": "one-shot", "provider": "spawn" } @@ -3884,11 +3883,10 @@ "config": { "provider": "deepseek-official", "model": "smoke-model", - "maxTokens": 256000, - "reasoningEffort": "high" + "reasoningEffort": "high", + "maxTokens": 256000 }, "adapterDefaults": { - "reasoningEffort": true, "maxTokens": true }, "system": "{{system}}", diff --git a/scripts/snapshots/python-sdk-single-exe/advanced/session.1.jsonl b/scripts/snapshots/python-sdk-single-exe/advanced/session.1.jsonl index db7a626752..39ac30a965 100644 --- a/scripts/snapshots/python-sdk-single-exe/advanced/session.1.jsonl +++ b/scripts/snapshots/python-sdk-single-exe/advanced/session.1.jsonl @@ -5,12 +5,12 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly DIRECT_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"{{messageId}}"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn","label":"Check direct child"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn","label":"Check direct child"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly DIRECT_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"{{messageId}}"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"{{messageId}}"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly DIRECT_CHILD_OK and","messageSeqs":[8],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"smoke-model","maxTokens":256000,"reasoningEffort":"high"},"adapterDefaults":{"reasoningEffort":true,"maxTokens":true},"system":"{{system}}","tools":["cordis_define","cordis_inspect_list","cordis_inspect_query","cordis_inspect_self","cordis_run","cordis_stop","cordis_undefine","job_kill","job_list","job_output","run_code","snapshot_double","subagent","workflow"]},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"smoke-model","reasoningEffort":"high","maxTokens":256000},"adapterDefaults":{"maxTokens":true},"system":"{{system}}","tools":["cordis_define","cordis_inspect_list","cordis_inspect_query","cordis_inspect_self","cordis_run","cordis_stop","cordis_undefine","job_kill","job_list","job_output","run_code","snapshot_double","subagent","workflow"]},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"smoke-model","contextWindow":1000000}} {"type":"session-log-deepseek/delivery-accepted","data":{"sessionId":"{{child-1}}","throughSeq":12}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} diff --git a/scripts/snapshots/python-sdk-single-exe/advanced/session.2.jsonl b/scripts/snapshots/python-sdk-single-exe/advanced/session.2.jsonl index eed0d049f5..7593678f03 100644 --- a/scripts/snapshots/python-sdk-single-exe/advanced/session.2.jsonl +++ b/scripts/snapshots/python-sdk-single-exe/advanced/session.2.jsonl @@ -5,12 +5,12 @@ {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"inserted":[{"content":[{"type":"text","text":"Reply with exactly WORKFLOW_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"{{messageId}}"}]}} {"type":"turn/start","data":{"turn":1}} {"type":"agent/inbox/spliced","data":{"target":"next-turn","start":0,"removedCount":1,"inserted":[]}} -{"type":"subagent/descriptor","data":{"version":2,"mode":"one-shot","provider":"spawn"}} +{"type":"subagent/descriptor","data":{"version":3,"mode":"one-shot","provider":"spawn"}} {"type":"step/start","data":{"turn":1,"step":1}} {"type":"user/message","data":{"content":[{"type":"text","text":"Reply with exactly WORKFLOW_CHILD_OK and nothing else."}],"source":{"kind":"user"},"role":"user","id":"{{messageId}}"},"surfaceOp":"append"} {"type":"user/message","data":{"content":[{"type":"text","text":"Current runtime context. This snapshot supersedes earlier runtime-context snapshots.\n\nCurrent DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations.\n\nApproval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`).\n\nYou are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}],"source":{"kind":"plugin","plugin":"@deepseek-ai/dsh-system-prompt","form":"snapshot","sections":[{"name":"sandbox:policy","text":"Current DSH file policy: danger-full-access. The DSH file sandbox does not restrict file modifications by available operations."},{"name":"approval:policy","text":"Approval prompts are disabled in this session: actions that require approval are rejected automatically — do not request sandbox escalation (do not set `sandbox_permissions`)."},{"name":"subagent:delegation","text":"You are a delegated subagent: your permission scope was fixed when you were started and cannot be widened from inside this session — operations that require approval are rejected automatically. When the task needs access beyond that scope, do not retry the denied operation; state the limitation in your reply so the delegating agent can handle it."}]},"role":"user","id":"{{messageId}}"},"surfaceOp":"append"} {"type":"session/title","data":{"title":"Reply with exactly WORKFLOW_CHILD_OK and","messageSeqs":[8],"source":{"kind":"fallback"}}} -{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"smoke-model","maxTokens":256000,"reasoningEffort":"high"},"adapterDefaults":{"reasoningEffort":true,"maxTokens":true},"system":"{{system}}","tools":["cordis_define","cordis_inspect_list","cordis_inspect_query","cordis_inspect_self","cordis_run","cordis_stop","cordis_undefine","job_kill","job_list","job_output","run_code","snapshot_double","subagent","workflow"]},"reason":"initial"}} +{"type":"request/header","data":{"header":{"config":{"provider":"deepseek-official","model":"smoke-model","reasoningEffort":"high","maxTokens":256000},"adapterDefaults":{"maxTokens":true},"system":"{{system}}","tools":["cordis_define","cordis_inspect_list","cordis_inspect_query","cordis_inspect_self","cordis_run","cordis_stop","cordis_undefine","job_kill","job_list","job_output","run_code","snapshot_double","subagent","workflow"]},"reason":"initial"}} {"type":"request/context","data":{"provider":"deepseek-official","model":"smoke-model","contextWindow":1000000}} {"type":"session-log-deepseek/delivery-accepted","data":{"sessionId":"{{child-2}}","throughSeq":12}} {"type":"assistant/chunk","data":{"turn":1,"step":1,"chunk":{"type":"block-start","index":0,"blockType":"text"}}} diff --git a/tsconfig.base.json b/tsconfig.base.json index 2c64da3ce6..8a8bc37baa 100644 --- a/tsconfig.base.json +++ b/tsconfig.base.json @@ -108,6 +108,7 @@ "@deepseek-ai/dsh-tools/presentation": ["./packages/core/tools/src/presentation.ts"], "@deepseek-ai/dsh-tools/types": ["./packages/core/tools/src/types.ts"], "@deepseek-ai/dsh-tool-subagent-control/list-agents": ["./packages/subagent/tool-subagent-control/src/list-agents.ts"], + "@deepseek-ai/dsh-tool-subagent/model-selection-settings": ["./packages/subagent/tool-subagent/src/model-selection-settings.ts"], "@deepseek-ai/dsh-user-approval/types": ["./packages/interaction/user-approval/src/types.ts"], "@deepseek-ai/dsh-user-questions/types": ["./packages/interaction/user-questions/src/types.ts"], "@deepseek-ai/dsh-agent/types": ["./packages/core/agent/src/types.ts"],