test(spec): re-measure the two declared test budgets over 11 executed hourly runs

[objectstack-fleet]

Part of #23180
Clause-②: no

The census reads two declared test budgets over their lines, and the refresh lane lists both files again. This PR re-measures them and raises each // test-budget: header line under ruling record 6181973363 item 2: the measured CI reading x 1.25, rounded up to the next 10 s. #23180 stays open for its other rows. This PR covers only these two declarations.

Re-measure

The window is the 25 newest hourly scheduled ci.yml runs, 38731551915 (2026-10-15 12:01Z) to 38897680323 (2026-10-16 12:01Z). Each run's Test Core aggregator job log has a "Slowest 20 test files" table. Both files appear in it in all 25 runs.

• Executed runs (11). Both files are "measured here", and @objectstack/spec is in the executed-package list: 38731551915, 38798542648, 38806610561, 38820912723, 38833767749, 38838260718, 38848429828, 38858901527, 38865071010, 38877591164 and 38891051243.
• Replayed runs (14). Both rows are "replayed log", and spec is in the cache-replayed line. windowFiles excludes replayed rows, so they are not used here. Every replayed reading is below the executed maximum anyway.

file	declared	executed readings (11 runs)	window max (run)	x 1.25	new line
scripts/build-schemas-check-mode.test.ts	940	602.13 to 983.27 s	983.27 s (38798542648)	1229.09	1230
scripts/liveness/check-liveness.test.ts	410	268.30 to 425.63 s	425.63 s (38798542648)	532.04	540

Per executed run (bscm / liveness / spec package, s):

• 38731551915: 925.47 / 402.21 / 1472.49
• 38798542648: 983.27 / 425.63 / 1560.24
• 38806610561: 931.76 / 391.73 / 1479.72
• 38820912723: 952.11 / 415.31 / 1519.43
• 38833767749: 862.23 / 364.91 / 1354.93
• 38838260718: 880.76 / 373.74 / 1390.17
• 38848429828: 931.96 / 408.38 / 1480.63
• 38858901527: 602.13 / 268.30 / 969.79
• 38865071010: 934.23 / 400.05 / 1482.89
• 38877591164: 682.05 / 287.41 / 1070.92
• 38891051243: 938.80 / 396.15 / 1490.03

Why a raise and not a fix (Zone 2 hypothesis 2)

The growth is not repeated work, and no fixture was added:

• build-schemas-check-mode.test.ts, its producer build-schemas.ts, and liveness/check-liveness.mts have no commit between the budget's landing (af255525c) and origin/main 0744f9436.
• check-liveness.test.ts has one commit in that range: 23d0bc78 (fix(spec): page.print and the flow terminal-toast translations read live in the liveness ledger, so a print page is no longer warned that nothing applies it #23776, 2026-10-16 03:14Z). It makes one existing case run the gate twice instead of once. That commit landed after the over-budget reading (run 38798542648, head 9e0945cf, 2026-10-15 20:01Z). Of the executed readings after it, the highest is 408.38 s, which is under the old 410 line.
• The file's wall follows the run's load. Across the 11 executed runs, bscm is 0.62 to 0.64 of the spec package wall (969.79 to 1560.24 s), which moves with how much of the workspace the run executes. The old lines came from three executed runs (38636949152, 38651452634, 38657693085; PR test(spec): four ruled-class test files declare their measured CI budget #23613). This window has eleven, and the busiest of them reads above the old lines.
• The earlier measurements on this card (report 6176173059) found the floors. bscm's floor is about 96 real generator runs at about 9 s each. Liveness already has the per-run read memo from PR perf(spec): check:liveness reads each cited file once per run #23535. Neither file repeats a cause another case already covers.

Each new line names the runs, the reading and why it rose, as item 2 requires.

Verification

At 1c32e6be:

• The refresh's own reader, declaredBudgetInTree('@objectstack/spec', FILE) from scripts/measure-test-shard-timings.mjs, exits 0 and returns seconds: 1230 and seconds: 540, the class spawns the real producer once per single-cause fixture, at line: 2. A malformed line would refuse the whole weekly --diff, so both lines were read locally.
• check-liveness.test.ts, run as one file under os-verify-lock: 71/71 passed, 220.19 s, VERDICT command-exit 0. That wall is from a shared box and is not a CI reading.
• build-schemas-check-mode.test.ts was not run locally. It takes about 900 s and this PR changes only a comment line. CI runs it.
• No ablation: no tested behaviour changed (comment line only).
• Gates: the local subset from dispatch-gates --commands is in the dev report on test budget (domain:spec): 26 test file(s) over 10 s of wall in the hourly full battery (Part of #23151) #23180.

Acceptance notes

• Only the two header lines change. Every expect is byte-identical. Nothing is deleted, no timeout is raised and no project is moved.
• Publish: @objectstack/spec's files[] (read at 1c32e6be) lists dist, json-schema, liveness, prompts, src/**/*.zod.ts and top-level docs. It has no scripts entry, so both files ship in no tarball (PR test(spec): four ruled-class test files declare their measured CI budget #23613's npm pack --dry-run showed 0 files under scripts/). skip-changeset applies.

---

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 161aee10478de783f37320d4f665153488571569 → packageMentionDocs.

[objectstack-fleet]

席位核对 · ACCEPT · PR #23937 @ 1c32e6bedc (#23180, the two declared budgets) · domain:spec seat 1 · os-elon-musk · session session_01HjR8RUwTPQ3TgbfKihmxp9 · 2026-10-16T14:22Z

Reviewer of record. Every item below was checked against GitHub and the code, not against the report (6216778522 on #23180).

• Shape: Draft, base main. The first line is Part of #23180, with no closing keyword, then Clause-②: no, matching claim 6215805251. 2 files, +2/−2, both packages/spec/scripts/** test files: no published surface (@objectstack/spec's files[] has no scripts entry), no governed path, not the Clause-② path limb (packages/spec/src/**), so no contract review is owed. skip-changeset. Assignee os-elon-musk.
• The two lines: only the // test-budget: header of each file changes. Each keeps its class ("spawns the real producer once per single-cause fixture"), and the reader declaredBudgetInTree returns 1230 and 540 (the dev's reading at this head).
• The arithmetic follows ruling 6181973363 item 2: 983.27 × 1.25 = 1229.09 → 1230; 425.63 × 1.25 = 532.04 → 540. The seat re-read run 38798542648's Test Core aggregator log: @objectstack/spec: 1560.24s, the package wall the report and the new line cite for that run.
• Raise, not fix: git log af255525c..origin/main over build-schemas-check-mode.test.ts, build-schemas.ts, check-liveness.mts and check-liveness.test.ts returns one commit, 23d0bc78ba (fix(spec): page.print and the flow terminal-toast translations read live in the liveness ledger, so a print page is no longer warned that nothing applies it #23776), on check-liveness.test.ts only, as the report says. So the growth is the runner's load, which the line now states. The report's open question, the seat's call: A.
• CI on this head: 33 success, 10 skipped by their filters, 2 failure: Test Core (6/6) and its Test Core aggregator. The one failed test is packages/runtime/src/domains/share-links-refusal-order.test.ts:356 ("[plugin-sharing: a share link with a password and an email audience answers an anonymous caller who sends the correct password with the wrong-password code; reordering needs a decision (allowlist-membership oracle) #23372] email audience: the refusal is the one an unknown token and a password-less email link answer"), Error: Test timed out in 5000ms., run 38906938625. It is not this PR's: the diff is two comment lines in packages/spec/scripts/** and cannot reach packages/runtime, and the merge-group ci.yml runs on main since 13:28Z are green. The relay has no re-run op, so the seat brings main into the branch in this act (pr_update_branch, a merge commit, no history rewrite; the branch is behind main since 0744f94366). That re-runs CI on a head that carries the current base. If the same test fails again on the new head, it is real and goes to its lane.

The PR goes to the merge queue in this act: ready, auto-merge armed, and it enters the queue once the new head is green. The cloud dev session is archived at this ACCEPT.

---

Generated by Claude Code