Files
deepseek-harness/packages/AGENTS.md
T
Yichen Jiang cc5173f4cf docs(testing): state the concurrent execution model where tests are written
The skill carries the reliability rules, but nothing an agent loads by default
said that specs run concurrently at all. The testing policy described tiers and
evidence without ever stating how a spec is executed, and neither subtree
AGENTS.md mentioned it — including scripts/, where the two suites that recently
failed on unrelated branches live.

docs/testing.md gains the execution model as the one home for the fact: forked
workers, concurrent coverage partitions beside other gates, and self-hosted
runners sharing a host and volume, with the rule that a spec passing only when
run alone is a defect in the spec. It links the skill for the detailed rules.

packages/AGENTS.md and scripts/AGENTS.md carry the short actionable form and
link that section, so the rule is present in the context loaded while a test in
either subtree is being written.

Both ceilings are raised for the added words and the targets in docs/AGENTS.md
move with them: docs/testing.md 1150 to 1300 (now 1237) and packages/AGENTS.md
675 to 750 (now 712), each keeping the 5% headroom the standard requires.
2026-08-29 15:31:12 +08:00

6.1 KiB

AGENTS.md — Harness Packages

These package-specific rules supplement the repo-wide conventions.

  • Plugin exports: service packages default-export their service class; function plugins named-export name / inject / Config / apply and have no default export. Mixing the forms makes the Loader discard the function plugin's namespace (postmortem).
  • Optional services use ctx.get(name). Reserve ctx.<name> for declared injections; the property proxy is topology-sensitive, while strict ctx.get reads the global service store (postmortem).
  • Product-visible plugins require a non-unit REAL-composition test. Hand-built ctx.plugin(...) suites are insufficient. Boot test-only cordis.yml through the Loader and app/process; mock only external services or nondeterministic inputs and assert model-visible, durable, or user-visible output. Keep opt-ins out of shipped defaults. Policy.
  • Initiator-owned private chains derive, then capture. Under ctx.agents.withInitiator(), recover the Agent at each orchestration entry, derive agent.session, and let operation-local helpers close over it. Keep Agent and Session explicit at lifecycle, session-log, service, authority, worker/process, persistence, and wire interfaces; do not widen a leaf helper from Session to Context merely to hide a parameter (rationale).
  • Represent one asynchronous operation with one lifecycle controller or transaction. Separate readiness, cancellation, disposal, reservation, or sentinel state requires an independent owner or settlement point; otherwise fold it while preserving rollback, callback containment, and quiescence.
  • Design Service Definitions for all current Consumers. Keep tool-schema, Loader, UI, transport, and provider-specific behavior in the Consumer or provider; do not let one Consumer dictate the service contract (capability-seam rationale). Inverse smell: a public service method with one internal caller — pass a private capability closure instead (RunCodeBridgeOptions).
  • Require a current owner and need. Tie each abstraction, state machine, option, defensive copy, and compatibility path to a current contract or production consumer, and keep behavior in its owning plugin or service.
  • Require evidence for public choices. Configurability does not justify an unsupported default, public operation set, format, or imported external concept. Use current-consumer evidence or relevant prior art; otherwise require an explicit value or defer the choice.
  • Write model-facing contracts from the model's perspective. Prompts, tool schemas, results, and diagnostics contain only task-relevant concepts, not UI, transport, or implementation vocabulary. Pin stable model-visible text verbatim and dynamic behavior through snapshots or end-to-end coverage.
  • Enforce a decision in the operation that makes it. Schema omission, prompt filtering, facades, wrappers, and listener order are not enforcement when direct or alternate callers can bypass them; test denial through the executor.
  • Publish state only at its commit point. Emit each notification and update derived state only after the operation succeeds; derive caches, prompts, UI echoes, replay, and query views from one authoritative source.
  • Apply bounds to the complete result. Enforce byte, token, item, and time limits where the complete emitted or retained value, including wrappers and metadata, is known; test tiny and exact limits, oversized single chunks, and multibyte byte limits.
  • Registry contributions prove disposal through the HMR-safety test required by testing policy: dispose the fiber and observe removal.
  • Specs run concurrently in forked workers beside other gate processes. Own each acquired port, path, and child process through teardown; a spec that passes only when run alone is a defect in the spec (execution model).
  • Every package owns ./invariant. Register the manifest name; check an event/data relation or give empty installers package-specific No runtime invariant: reasons. Generated companions, unexplained empties, and ignored reporters fail verify-package-invariants.

Naming rules:

  • Package tsconfig: extends tsconfig.base.json (Client: tsconfig.base.client.json), sets rootDir: src and outDir: lib/types, references workspace dependencies plus runtime-diagnostics/invariants, and registers in one aggregate. Packages with distinct Host and Client compiler faces use tsconfig.host.json and tsconfig.client.json leaves plus a solution-only root; ordinary two-entry Client plugins do not split (layout).
  • src/types.ts contains only types — no runtime code.
  • Tests live at package level under tests/, not src/__tests__/.
  • Update package README and JSDoc contracts in the same commit as behavior, and verify them against code with dsh-prose-standard. Group READMEs declare subsystem ownership through a canonical English page link or justified exemption.
  • Package READMEs document model, token, and KV-cache effects using the canonical Model Experience format.
  • Package READMEs put durable consumer gaps and non-obvious maintainer constraints under ## Known Limitations and Deferred Work; ordinary cleanup stays in its TODO or Agent Note. Packages with none use a justified allowlist entry (rationale).