[objectstack-fleet]

Requested by Leehom · project thread
Part of #23181 (slice 9). The card stays open for its other rows.

Clause-②: no
Metadata-protocol: none

Test files only, so this PR carries skip-changeset: both files are *.test.ts under packages/verify/src, which the package's build does not ship.

What changed

Two @objectstack/verify harness suites booted a fresh in-process stack in every case and paid the boot's first load inside the first case's clock.

• src/harness.app-default-profile.test.ts: await warmUpBoot() at module scope (boot-warm-up.testkit.ts, the convention slices 3 and 4 used). The first two cases now read ONE boot of the declared-default configuration from a beforeAll. The first reads the profile the boot wired; the second signs a member up and reads as them. Neither reads what the other writes. Boots: 4 → 3.
• src/harness.org-context.test.ts: await warmUpBoot() at module scope. One beforeAll boots app with the flag off, records what the cases read, and stops. Then it boots with orgContext: true, records, and stops, in the same order the ruling case always booted them. The ruling case compares the two recorded postures, the bind case reads the orgContext: true boot, and the default-boot case reads the default boot. Boots: 4 → 2.

Every assertion reads the same value of the same boot shape (same configuration, same options) as before. The test full names are identical before and after (sorted fullName lists compared).

The reading, and the rows this PR brings under 10 s

The reading is refresh PR #25960's per-file list (runs 40200279306 and others in its window). The newest LANDED refresh, #25409, lists no per-file rows; slice 8's report records why. Of this family, #25960 still lists five rows:

file	#25960 slowest (runs over / executed)	local before → after, median of 3	this PR
src/harness.app-default-profile.test.ts	18.90 s (1 / 2)	6608 → 3126 ms (×0.47)	under 10 s: projected 8.9 s in CI
src/harness.org-context.test.ts	16.60 s (1 / 2)	6741 → 2431 ms (×0.36)	under 10 s: projected 6.0 s in CI
src/harness.key-custody.test.ts	22.41 s (2 / 2)	6852–7338 ms, untouched	open: see Acceptance notes
src/harness.one-composition.test.ts	17.89 s (2 / 2)	9206–10354 ms, untouched	open: see Acceptance notes
src/handle.test.ts	17.30 s (1 / 2)	3857–4253 ms, untouched	open: already warmed, one shared boot

The pin is the next landed refresh: it should no longer name the two rows marked under 10 s.

Measurement

On one 4-vCPU box, under os-verify-lock. It is a shared box, so the ratios carry the result rather than the absolute times. Each run is one vitest run --maxWorkers=2 process per file. The number is the default reporter's file line (the module's duration: hooks included, collection excluded), which is the figure the timing refresh reads. The json reporter's endTime - startTime leaves out a beforeAll that runs before the first case, so it is not used. BASE a6e8f95d6 copies and HEAD files were interleaved A/B in the same lock hold.

run	app-default BASE	app-default HEAD	org-context BASE	org-context HEAD
1	6488	3126	6523	2431
2	7591	3523	6741	2702
3	6608	3023	7223	2263

Cases still isolated. vitest run --sequence.shuffle --sequence.seed=N over both files, seeds 1 to 5: 11 / 11 passed each time. With seed 1, the member sign-up case ran before the profile read on the shared boot. With seed 2, the default-boot case ran before the ruling case.

Where a boot's time goes. A CPU profile of six warm bootStack calls (0.82–1.09 s each) puts the largest share in the default datasource's schema sync on sql.js (knex ensureConnection / query → driver-sqlite-wasm _query, driver-sql initObjects). A test file can only boot fewer times; it cannot make a boot cheaper.

Acceptance notes

• harness.key-custody.test.ts (22.41 s, 2 / 2) is untouched. Each of its 5 boots runs under its own key posture, so no two boots can be shared. A warm-up is excluded there by slice 4's seat ruling (option A on 6221027193): it would build the harness data key outside the keyless posture the file inspects.
• harness.one-composition.test.ts (17.89 s, 2 / 2) is untouched. It is already warmed. Each of its 12 boots is the subject of its case (the instance rule and composition), so no two can be shared. It reads 9.2–10.4 s locally.
• handle.test.ts (17.30 s, 1 of 2 runs over) is untouched. It is already warmed and boots once per file. The one reading over 10 s came from the window's most loaded run (40200279306, where all five rows took their slowest).
• #23181 remains open for those rows and for the other packages' rows.

Tests

• pnpm --filter @objectstack/verify typecheck: VERDICT command-exit 0 (check:test-typecheck: OK, 0 debt).
• eslint --no-inline-config packages/verify: exit 0.
• Both files: 11 / 11 passed in 6 A/B runs and 5 shuffled runs.
• @objectstack/verify whole suite (vitest run --maxWorkers=2): Test Files 46 passed (46), Tests 393 passed (393), VERDICT command-exit 0.
• Gates at cb29c5b4d: dispatch-gates --commands derived 55 commands, and all 55 exit 0. --ran reports "55 derived, 55 run, 0 NOT-MEASURED, 0 UNRUN" and exits 0. On the first pass two gates exited 3 with PREREQUISITE NOT MET, which is not a verdict: check-plugin-teardown-shape --self-test (shallow clone; fixed by fetching 621a4876) and check:dual-build-cjs-loads (eight packages had no dist/; fixed by building them). Both exit 0 on the re-run.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

Nothing in this diff resolved to a documentable surface (no symbol, route or SDK anchor derived from 0 changed package(s)), so this run has no opinion about the docs.

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 — 0 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.

[objectstack-fleet]

席位核对 · ACCEPT · domain:server seat 1 · os-project-manager · session_01R1q2xdtYgU6FKrZ7LgjPCC · 2026-10-26T15:13Z

Head cb29c5b4df, report 6419000068 (on #23181, slice 9), claim 6417592580 and its amendments. No contract review owed: test-only, no packages/spec/**.

item	reading
Shape	Draft, base main, Part of #23181 (slice 9), skip-changeset, mergeable clean.
Files	2 files, +99/−63: packages/verify/src/harness.app-default-profile.test.ts and harness.org-context.test.ts. Both use the existing boot-warm-up.testkit.ts (on main).
Fix	Each file pays the boot's first load at collection (warmUpBoot()), and cases that read the same boot configuration share one boot in a beforeAll. app-default-profile: the two read-only cases on the declared default. org-context: one bootAndRead per option set, then the cases assert its reading. Every case keeps its assertion. No timeout raised and nothing deleted.
Evidence	app-default-profile 6,608 → 3,126 ms and org-context 6,741 → 2,431 ms (interleaved A/B, median of 3); projected CI 8.9 s and 6.0 s. Shuffle seeds 1–5 green. The verify suite passes 393/393. 55/55 gates exit 0, --ran reconciled. On this head 36 check runs succeeded and 16 skipped as expected, with 0 failed.
Carried	key-custody, one-composition and handle.test are left untouched, with reasons given in the report. The open question goes with the dev's recommendation: wait for the next refresh, then file a card on bootStack's per-boot cost if those rows remain.
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