chore(pm): tighten pm-dispatch SKILL.md table-row pin to 642

[zhuangjianguo]

Fixes#13090

What

Follows the ratchet gate's own standing hint: lowers MAX_TABLE_ROW_BYTES for
.claude/skills/pm-dispatch/SKILL.md in scripts/pm/check-skill-line-ratchet.mjs
from 765 to 642, matching the widest table row after PR #13085 (MERGED
2026-08-29T09:35:17Z) shrank the domain:engine row. Gate-script edit only —
.claude/skills/pm-dispatch/SKILL.md itself is untouched. No other pin was
touched (this is the only pin with paid-down headroom this round).

Proof (headroom returns to 0, as expected)

Before (pre-edit, at merge-base 56c5b1d):

ℹ️ .claude/skills/pm-dispatch/SKILL.md: max-table-row-bytes headroom is 123 — every pin is
seeded at 0 headroom, so a row has been paid down; lower the pin to 642
(shrink-only ratchets tighten opportunistically).
✓ check-skill-line-ratchet: .claude/skills/pm-dispatch/SKILL.md: widest table row is 642 bytes
(pin 765; headroom 123).

After (at head d4c5141b0):

✓ check-skill-line-ratchet: .claude/skills/pm-dispatch/SKILL.md: widest table row is 642 bytes
(pin 642; headroom 0).

No hint line prints post-edit — headroom is 0, which is the expected/intended
result of a freshly-paid-down pin, not a failure.

node scripts/pm/check-skill-line-ratchet.mjs — EXIT 0
node scripts/pm/check-skill-line-ratchet.mjs --self-test — 111 cases pass, EXIT 0
pnpm check:pm-skill-ratchet — EXIT 0 (runs both of the above)

Local gates run at head d4c5141b0

Derived from the actual diff via node scripts/pm/dispatch-gates.mjs (not copied
from the dispatch brief) — all green:

• pnpm check:agent-test-spelling
• pnpm check:bash32-floor
• pnpm check:cli-command-ids
• pnpm check:cross-package-test-inputs
• pnpm check:entry-guard
• pnpm check:parse-guard
• pnpm check:pm-skill-ratchet
• pnpm check:pnpm-filter-targets
• pnpm check:watch-hint-literal
• node scripts/check-ci-filter-parity.mjs
• node scripts/check-cross-package-test-inputs.mjs
• node scripts/check-shard-attestation.mjs

Plus two convention-triggered obligations the derivation surfaces for any
gate-SCRIPT edit (not named in the dispatch brief; re-derived and run here):

• node scripts/pm/bare-root-worklist.mjs --self-test — OK, 50 live rows, none stale/missing/contradicted
• pnpm check:pm-dispatch-gates — dispatch-gates self-test: 897 cases pass

node scripts/check-test-completeness.mjs run bare (no turbo log) exits 3 —
per its own printed text this is PREREQUISITE NOT MET / NOT MEASURED, not a
finding; CI supplies the log itself.

Scope

scripts/pm/check-skill-line-ratchet.mjs only — one line. No separate test
file exists for this script (its self-test lives inline via --self-test),
so no additional test-file surface to touch. skip-changeset applies (closed
list: scripts/pm/**, no package release) — label applied on open.

---

Generated by Claude Code

[github-actions]

⛔ merge queue 构建失败 — 先分诊，再决定要不要重排

队列构建 33298233173 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集），
所以失败的测试可能在本 PR 没碰过的包里 —— 那不是重排能修的。每次盲目重排都会让排在后面的所有 PR 重建一轮。

失败的 job（日志抽取，best effort）：

• Test Core (5/6) — 失败步骤: Run this shard's tests

  @objectstack/rest:test: FAIL src/ui-view-route-tenancy.measurement.test.ts > [#13214] §3 what the crossed response actually contains > ⚠️ `hidden` is NOT a uniform floor on the crossed path — measur
  ↳ 失败原因: @objectstack/rest:test: AssertionError: expected [ 'name', 'created_at', …(1) ] to include 'status'
  

↳ 失败原因 是判读的关键：超时（Test timed out in … / Hook timed out in …）多半是负载/时序，不是本 PR 的回归；
断言（AssertionError: …）才指向真实的行为改变。两者的 FAIL 行长得一模一样，只有这一行能区分。

跨 PR 相同签名（24h，按失败测试文件聚合）：

• src/ui-view-route-tenancy.measurement.test.ts — 24h 内已弹出 4 个 PR（test(rest): drive the tenancy axis of GET /ui/view/:object/:type — an anonymous caller can name another environment and receive its UI view #13258、fix(driver-memory): a no-value row satisfies $nin / $notContains in the reference matcher (#13166) #13356、docs(spec): tool.json's parameters note records the post-#13317 reality #13359、chore(pm): tighten pm-dispatch SKILL.md table-row pin to 642 #13360），但它们同属 1 条投机栈 ⇒ 1 次独立命中。汇总 issue：Queue-flake anchor: src/ui-view-route-tenancy.measurement.test.ts #13374（已刷新）
• ⚠️ 24h 评论账本没读完（超过 5 页仍未读到窗口尽头），所以上面的「不同 PR 数」是下界，不是全量。

历史信号：

• 本 PR 过去 24h 无队列失败记录（首次）。
• 过去 24h 队列共有 13 个失败构建（不含本次）。

分诊清单：

1. 失败测试在本 PR 改动的包里 → 真回归，修 PR。
2. 失败测试与本 PR 无关 → 看上面的「跨 PR 相同签名」；已有汇总 issue ⇒ flaky/环境问题实锤，去那张 issue 上谈，修好前重排只会再烧一轮全队列。
3. 两者都不是 → 可能与同组 PR 语义冲突；等前面的 PR 落地或失败出队后再重排一次即可，不要连续重排。

---

Generated by Claude Code · merge-queue-triage workflow (#4859)

[zhuangjianguo]

Queue-ejection disposition (skills lane seat, session session_01EXxTW8mvPBhoHxmyPZ63de): the failing check is queue build 33298233173, Test Core (5/6), signature src/ui-view-route-tenancy.measurement.test.ts — not this PR's failure. This diff is one line in scripts/pm/check-skill-line-ratchet.mjs; the failing test file exists only on PR #13258's branch and reached this build through speculative stacking (ledger #13374 marks this PR S1 · inherited, a bystander by construction). Root cause is already established on the ledger by the domain:engine seat: the assertion pins pre-fix behaviour that merged fix PR #13329 (2026-08-30T04:23Z) reversed by design — deterministic, not flaky; the fix is an in-place assertion inversion inside #13258, owned by that PR's author (told there before enqueue).

Disposition per the triage checklist's own branch 3 and the one-re-run rule: the stack root has exited the queue, this PR's own tree is current-main + one line, so it re-queues once now. If it is ejected again on the same signature it will NOT be re-queued again — it then waits explicitly on #13258's assertion flip, and that wait will be stated here, not silent.

---

Generated by Claude Code