[objectstack-fleet]

Requested by Leehom · project thread

Part of #23181 (slice 17). The card is a series and stays open.

Clause-②: no
Metadata-protocol: none

Test-only, comment-only diff: skip-changeset. Eight lines under packages/cli/test/, which is not in @objectstack/cli's files[] (dist, README.md, CHANGELOG.md); no files[] changes in any package.

维护者速读

• 改了什么:给 cli 里 8 个真实启动 os verify 的测试文件(test/verify-*.test.ts)各加一行第 2 行的 // test-budget: 声明,只是注释。
• 为什么改:这 8 个文件在 CI 上仍超 10 s;逐个数过,每个文件里每次 CLI 启动都是不同的(命令、参数、夹具)组合,没有重复可共享,而且已在 beforeAll 里各跑一次,所以按裁决 6181973363 B 声明实测预算。
• 风险与回滚:零运行时风险,不改断言、超时与分层;回滚即删这 8 行。
• 先看:packages/cli/test/verify-parent-write-reach.test.ts、packages/cli/test/verify-reference-shapes.test.ts 第 2 行(预算最大的一行,和第三次 CI 读数抬高了的一行)。
• 两分钟验证:用 scripts/measure-test-shard-timings.mjs 的 readTestBudget 解析这 8 个文件,预期 8 行都返回 spawns the real producer once per single-cause fixture、行号 2。
• 需要你做的:无,席位落地。

What changes

One header line at line 2 of each file below, in the ruled grammar // test-budget: N — spawns the real producer once per single-cause fixture: REASON (ruling record 6181973363, letter B; grammar readTestBudget in scripts/measure-test-shard-timings.mjs). N is the slowest CI reading x 1.25, rounded up to the next 10 s. Nothing else changes: no assertion, timeout, tier or helper.

file	tests	CLI spawns (counted)	distinct (command, flags, fixture)	CI 40392443783 (s)	CI 40424746941 (s)	CI 40447801748 (s)	local (s)	budget
test/verify-parent-write-reach.test.ts	3	3, in one beforeAll	3	55.76	36.47	35.05	40.9	70
test/verify-delete-behaviors.test.ts	3	3, in one beforeAll	3	52.68	34.05	31.86	35.6	70
test/verify-dashboard-windows.test.ts	3	3, in one beforeAll	3	50.49	34.23	32.75	35.3	70
test/verify-unique-scopes.test.ts	3	3, in one beforeAll	3	48.96	49.91	31.63	36.1	70
test/verify-reference-shapes.test.ts	3	3, in one beforeAll	3	37.07	43.50	45.37	32.1	60
test/verify-author-time-stage.test.ts	3	4, one per case and step	4	38.85	28.55	31.82	31.0	50
test/verify-json-stdout.test.ts	4	2, in one beforeAll	2	32.04	32.30	21.23	21.3	50
test/verify-host-root.test.ts	2	1, in its beforeAll	1	13.13	9.64	10.37	10.5	20

Why

The count. In the five flag files (parent-write-reach, delete-behaviors, dashboard-windows, unique-scopes, reference-shapes) the three spawns are verify --json on the app, verify --FLAG --json on the app, and verify --FLAG --json on the control fixture (in reference-shapes, the same directory under OS_ALLOW_LAX_VALUE_SHAPES=1). author-time-stage runs validate --json, verify --json and verify on the planted stack and verify on the clean control. json-stdout runs verify --json and verify once each. host-root runs one verify --app from a foreign cwd. No triple repeats inside a file, and every fixture is the file's own, so there is no spawn to share. Seven of the eight already spawn in one beforeAll; author-time-stage spawns inside its cases, one distinct triple each, so hoisting them would move the cost and save nothing. What is left is the spawn floor.

The readings. No shard-timings refresh has run on main since slice 14 (PR #26256). The CI numbers are the vitest module lines (they include beforeAll) of the three scheduled full-battery ci.yml runs on main after it: slice 15's two, 40392443783 and 40424746941, and the newer 40447801748 (head 26e7c75e9a; its Test Core shards and aggregator all passed, only a Test UI shard failed). The newer run raises one file's maximum: verify-reference-shapes 45.37 s, which still rounds to 60.

No file left out. verify-host-root is under 10 s in one reading (9.64 s) but over in the other two, so it takes a header. No file needed another mechanism.

Tests

Comment-only diff, so behaviour and wall are the tree's. A/B of a comment line measures nothing (slice 15's precedent), so the after reading is one run on this head's tree.

Local run of the family (one lock hold)

cd packages/cli && pnpm exec vitest run --project integration --maxWorkers=2 --reporter=default plus the eight files, under os-verify-lock, on a shared container, manifest setup on (queue tier): Test Files 8 passed (8), Tests 24 passed (24), VERDICT command-exit 0, lock held 147 s. Every file reads under its declared budget.

Header parse

All eight lines parse through readTestBudget (exit 0): each returns class spawns the real producer once per single-cause fixture at line 2, seconds 70 / 70 / 70 / 70 / 60 / 50 / 50 / 20.

Typecheck and lint

pnpm --filter @objectstack/cli typecheck: VERDICT command-exit 0 (check:test-typecheck: OK). pnpm exec eslint --no-inline-config on the eight files: exit 0. pnpm check:nul-bytes: exit 0.

Acceptance notes

• Pin: the next shard-timings refresh lists none of these eight files in its over-budget table. Each comes back only if a run reads it over its own budget.
• The budgets rest on three CI readings. A refresh over its 20-run window may read one of them higher, and that is the refresh's job.
• packages/verify's own files over 10 s in slice 15's local reading were measured only, not changed; their causes are in the slice report on test budget (domain:cli): 119 test file(s) over 10 s of wall in the hourly full battery (Part of #23151) #23181 for a later slice.

Verification

Local gates at dfec98106: node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack derived 48 commands. Each ran with its exit code captured before any pipe. 46 exited 0 on the first pass. check-plugin-teardown-shape.mjs --self-test exited 3 on the shallow clone and 0 after git fetch origin 621a487607881c66b2899b7e3477115229a156b4. check:dual-build-cjs-loads exited 3 (ten packages outside the cli closure had no dist/) and 0 after building them. --ran reconciliation: 48 derived famil(ies) accounted for — 48 run, 0 NOT-MEASURED, exit 0. The repo-wide pnpm lint is left to CI.

Human Minimum

• 1. Schemas start as Zod; packages/spec holds no runtime logic. (not touched)
• 2. Every name I added uses the casing for its kind. (no names added)
• 3. No formatter ran; pnpm lint passes. (eslint on the touched files: exit 0)
• 4. One issue, one branch off an up-to-date main, one pull request.
• 5. A change that publishes carries a changeset, and both declaration lines are filled in here and in the changeset. (publishes nothing: skip-changeset)
• 6. packages/spec touched: its generated artifacts are regenerated and committed. (not touched)
• 7. No hand edits to content/docs/references/, packages/*/CHANGELOG.md or content/docs/releases/.
• 8. Title, body and comments are in English. (the maintainer brief is in Chinese, per AGENTS.md)
• The Before Submitting list in CONTRIBUTING.md is done.

🤖 Generated with Claude Code

https://claude.ai/code/session_01FEkQ5nuFC67LujGL3Fsyiy

[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 98c025775df1b77449e89465dbd2469f85dc16a1 → packageMentionDocs.

[objectstack-fleet]

席位核对 · ACCEPT · domain:server seat 1 · os-project-manager · session_01R1q2xdtYgU6FKrZ7LgjPCC · 2026-10-27T06:09Z

Head dfec981068, report 6434906896 (on #23181), claim 6434344323. No contract review is owed: there is no packages/spec/** diff, no governed path, no narrowing and no Clause-②: yes.

item	reading
Shape	Draft, base main, Part of #23181 (slice 17), skip-changeset (test-only, no files[] change). 8 files, +8/−0: one comment line each, all in the claimed os verify family.
Fix	All eight test/verify-*.test.ts files spawn only distinct (command, flags, fixture) triples, so each gets the ruled // test-budget: header at line 2 (ruling 6181973363 letter B) and nothing else.
Seat check	The seat recomputed all 8 budgets from the slowest of three scheduled CI readings × 1.25, rounded up to 10 s: 55.76→70, 52.68→70, 50.49→70, 49.91→70, 45.37→60, 38.85→50, 32.30→50, 13.13→20. No assertion, timeout or project placement changes.
Evidence	CI on head dfec981068 is green: 36 checks succeeded, 16 were expected skips, and none failed or are pending; mergeable clean. Locally 8 of 8 files and 24 of 24 tests; every header parses through readTestBudget. Derived gates: 48 of 48 run, with 0 NOT-MEASURED.
Carried	packages/verify's six over-10 s files were measured only, with each one's cause in the report, for a later slice. Slice 16 (#26493) is in the merge queue on other files.
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, once every check on this head is green or an expected skip.

---

Generated by Claude Code