[objectstack-fleet]

Requested by Leehom · project thread

Fixes #26047

Clause-②: no
Metadata-protocol: none

What was wrong

The merge queue ejected three independent PRs (#25979, #26007, #26026) on the same line: hook-timeout-override-refusal.test.ts › ⭐ CONTROL, AssertionError: expected 1 to be +0. The vitest list child exited 1. Load was not the cause.

packages/cli/vitest.config.ts builds its tier populations at config load through packages/cli/vitest-tiers.ts. testFilesOnDisk walked the whole package, including packages/cli/tmp/, which is git-ignored scratch. Sibling tests in the same slice render whole scaffolds into tmp/ and delete them mid-run (init-template-comments-self-contained, scaffold-manifest-schema). Those scaffolds hold real test files such as package.test.ts and src/index.test.ts. When a delete landed inside the walk, the config load threw ENOENT (scandir or open), so the child exited 1. Two other effects of the same walk: a scaffold's test files joined the unit population while they existed, and the outer vitest run config load could die the same way.

Change

• packages/cli/vitest-tiers.ts: tmp added to SKIP_DIRS, with a comment saying why. This is the seat's ruling A, recorded as a claim amendment on Queue-flake anchor: test/hook-timeout-override-refusal.test.ts #26047.
• packages/cli/test/hook-timeout-override-refusal.test.ts:
  • Every status assertion now carries the child's exit status and the first 2000 characters of its output. The next red says why.
  • The no-override control plants planted.test.ts under a fresh tmp/ directory and asserts the listing does not name it. That is the red-first pin.
  • No timeout raised, no skip, no retry. The [finding] in packages/cli the --hookTimeout CLI override is SILENTLY INERT under test.projects — the instrument the repo's own prior art uses to WITNESS a cold load leaving the clocked window returns green having measured nothing #18788 verdicts are unchanged: --hookTimeout and --teardownTimeout refuse, --testTimeout passes, and the no-override control passes and says nothing.

Evidence

All runs on a shared 4-core container, with the @objectstack/cli dependency closure built.

Reading	main 0a69482	this PR 358e11b
The control's exact child (vitest list --filesOnly, no override) looped beside the real scaffold-rendering tests	3 of 132 exit 1, each failed to load config … ENOENT … packages/cli/tmp/manifest-init-plugin-…/package.test.ts or …/render-create-plugin-…/src/index.test.ts	0 of 216 (120 + 96)
Test file under synthetic tmp/ churn	red on the first run (vitest list exited 1; its output began: failed to load config … ENOENT … scandir …)	0 of 8 red
Test file alone beside 8 busy loops (load, no churn), 5 runs	5 of 5 green; file 27.5–31.3 s, control case 6.7–7.7 s	5 of 5 green; file 27.6–30.9 s, control case 6.4–7.6 s
Population (vitest list --filesOnly)	506 files	506 files
Red-first pin against main's walk	× ⭐ CONTROL …: the listing named [unit] tmp/hook-timeout-control-Btvz04/planted.test.ts	green

The load-only row rules out the time hypothesis: the control child takes about 7 s against its 27 s budget. A directory watcher across all three CI slices (OS_TEST_SHARD k/3) recorded that the only directories appearing or vanishing in the package tree were under tmp/. The slice holding this test churned 25 of them. CI job logs are not readable from this container, so the queue's own stderr was not seen. The new message makes the next queue red carry it.

Tests and gates (head 358e11b)

• pnpm --filter @objectstack/cli exec vitest run --project unit: 333 files / 4854 tests passed. The integration tier is left to CI, because this diff touches no spawn entry.
• pnpm --filter @objectstack/cli run typecheck: exit 0.
• eslint packages/cli --no-inline-config: exit 0.
• dispatch-gates --commands: 55 derived, 54 run with exit 0, 1 NOT MEASURED (check:dual-build-cjs-loads, which needs a full build and exited 3). --ran reconciles: 0 unrun. This includes check:tier-file-adoption and check:cross-package-test-inputs, both green, and neither reads the cli SKIP_DIRS. check-plugin-teardown-shape --self-test was rerun green after fetching its positive-control commit. Its first attempt overlapped my synthetic tmp/ churn and refused on an unreadable tmp/ file, which is the same walk class (see Acceptance notes).

Changeset

skip-changeset: @objectstack/cli publishes files: ["dist", "README.md", "CHANGELOG.md"], and neither touched file is in it or built into dist.

Acceptance notes

• Same walk shape, not changed here, not measured as a failure in CI: packages/qa/vitest-filter-preflight/src/index.ts testFilesUnder (filtered runs only), scripts/nightly-tiers.mjs tierPackages, and check-plugin-teardown-shape all descend into packages/cli/tmp/. A churn in tmp/ during one of them can refuse on a vanished file. Carrier: none.

🤖 Generated with Claude Code

https://claude.ai/code/session_01REwoooZ6nXESkVCEtsbeRC

[github-actions]

📓 Docs Drift Check

This PR changes 1 package(s): @objectstack/cli, touching 2 documentable anchor(s).

5 hand-written doc(s) NAME something this change touched and may need an implementation-accuracy re-verification:

• content/docs/ai/skills-reference.mdx (via node_modules (literal, a string literal in SKIP_DIRS))
• content/docs/ai/skills.mdx (via node_modules (literal, a string literal in SKIP_DIRS))
• content/docs/deployment/cli.mdx (via node_modules (literal, a string literal in SKIP_DIRS))
• content/docs/deployment/self-hosting.mdx (via node_modules (literal, a string literal in SKIP_DIRS))
• content/docs/upgrading.mdx (via node_modules (literal, a string literal in SKIP_DIRS))

⛔ 1 release-owned page(s) also name something this change touched. These are read-only:

• content/docs/releases/v18/18-0.mdx (via node_modules (literal, a string literal in SKIP_DIRS))

content/docs/releases/ is RELEASE-OWNED (AGENTS.md "Documentation Guardrails"): release
notes are written centrally at release time, and a code PR that edits them is the exact PR
that guardrail exists to stop. They are still audited — read-only. If one of them is actually
wrong, file an issue or open a dedicated docs-only PR; do not edit it here.

What this run could not see

• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.
• a key NAME is not a key, so the hand re-read the line above prescribes can land on the wrong schema. The same spelling is authorable on one governed type and a [REMOVED] tombstone on another for each of active, aria, joins, objects, template, tools and version (censused on [finding] tools is a key on BOTH AgentSchema (tombstoned, dead) and SkillSchema (live, cloud-attested), so a name-based search attributes skill examples to the agent key — it produced a false stop-the-line alarm on PR #19059 #19093 over the liveness ledger's governed types, top-level keys); nothing in a search result distinguishes the two, so a grep hit on a LIVE example reads as evidence about the DEAD key. Measured on fix(spec): the agent.tools liveness row says dead — it claimed live on a key the schema tombstoned #19059: content/docs/ai/agents.mdx was reported as contradicting the agent.tools tombstone over its tools: example at :161, which is inside the defineSkill({ block opened at :155 — the page was already correct. Settle ownership by PARSING the value against both schemas, never by the name: that literal PASSES SkillSchema, and as an AgentSchema it FAILS at tools with the tombstone prescription. ⛔ These names are not the whole class — a key retired through a .strict() guidance map leaves no tombstone in the walked shape and none of them here (tool.category, live as AIToolDefinition.category).

Coarse fallback — 31 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json c7aa9cde24ca7d17563caa4e01037f3ab1b9e990 → packageMentionDocs.

Which tree this was computed on

This run read content/docs from 9973670cec23c77457a7e18fb94b4b46098a5008 — the merge of head 358e11b89c06ff37f56d1cff069aa750f198c05e into base c7aa9cde24ca7d17563caa4e01037f3ab1b9e990, which is what actions/checkout gives a pull_request run. Not the PR head.

A worktree cut from an older main holds a different content/docs, so re-deriving there can legitimately return a different list — that is a different tree, not a wrong row. To answer on the same tree:

# while this PR is open — GitHub drops the merge commit once it closes
git fetch origin 9973670cec23c77457a7e18fb94b4b46098a5008 && git checkout 9973670cec23c77457a7e18fb94b4b46098a5008
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin c7aa9cde24ca7d17563caa4e01037f3ab1b9e990 358e11b89c06ff37f56d1cff069aa750f198c05e && git checkout -B drift-repro c7aa9cde24ca7d17563caa4e01037f3ab1b9e990 && git merge --no-ff 358e11b89c06ff37f56d1cff069aa750f198c05e

node scripts/docs-audit/affected-docs.mjs --json c7aa9cde24ca7d17563caa4e01037f3ab1b9e990

⚠️ That checkout carried uncommitted changes, so the commit above does not fully identify what was read.

Advisory only, and a precision-first one (#9192): a page is listed because it names a
symbol, wire route or SDK method this diff touched — not because it mentions a changed
package. Each row says which anchor put it there, so a wrong row is reportable rather than
merely annoying. To re-verify, run the docs-accuracy-audit workflow scoped to these files:
node scripts/docs-audit/affected-docs.mjs c7aa9cde24ca7d17563caa4e01037f3ab1b9e990 → pass the list as
args.docs, on the commit named under Which tree this was computed on.

[objectstack-fleet]

席位核对 · ACCEPT · domain:server seat 1 · os-project-manager · session_01R1q2xdtYgU6FKrZ7LgjPCC · 2026-10-26T15:11Z

Head 358e11b89; report 6419032521 (on #26047); claim 6415802873; amendment 6417872008 (seat ruling A). No contract review is owed: there is no packages/spec/** diff and no narrowing.

item	reading
Shape	draft, base main, Fixes #26047, skip-changeset. vitest-tiers.ts is outside @objectstack/cli files[] (dist, README.md, CHANGELOG.md), as the body states. Mergeable clean.
Files	2 files, +52/−7: packages/cli/vitest-tiers.ts adds tmp to SKIP_DIRS with a comment; packages/cli/test/hook-timeout-override-refusal.test.ts changes the pin.
Root cause	Not load: 5/5 green under 8 busy loops. The config-load walk read the git-ignored packages/cli/tmp, where sibling tests render scaffolds (with real *.test.ts files) and delete them. A delete mid-walk threw ENOENT out of the config load, so the ⭐ CONTROL child exited 1. Reproduced on main in 3 of 132 runs with the real neighbours.
Fix	Ruling A, one entry in SKIP_DIRS. The pin's status assertions now carry the child's exit and output, so the next red says why. The control plants tmp/.../planted.test.ts and asserts it is not listed (red against main's walk).
Evidence	0 of 216 on the fix (main 3 of 132). Population unchanged at 506; check-tier-file-adoption and the cross-package test-input census green. 36 check runs success and 16 expected skips on this head, 0 failed. Build Core (hosting check:dual-build-cjs-loads, the one gate not measured locally) is green. cli unit tier 333 files / 4,854 tests green.
Landing	「按照 pm-dispatch 规则复核过的都授权」 (Leehom, 2026-10-25T11:35Z): ready and auto-merge through the relay.

---

Generated by Claude Code