[PM seat] skills — 🟢 os-zhuang

[huangyiirene]

本贴是 skills 座位的唯一权威登记。 座位贴协议:索引 = label:pm:seat,总入口 #4604。六段模板(#7583):范围 | 当前 PM | 继承台账 | 热文件串行队 | 常设承诺 | 说明 —— 段段现值,⛔ 不写班次叙事。

范围

Both skill roots: .claude/skills/** + skills/**. Routing label domain:skills. (Also .claude/agents/os-dev.md / AGENTS.md instruction-architecture work per the 2026-08-12 approvals — conservative ADR-class treatment.)

当前 PM

🟢 os-zhuang — session session_012Gg6rMAti8ZaueWb6BRsDn, took over 2026-08-12 ~01:5xZ. Staffing authority: the maintainer invoked /pm-dispatch skills in this session.

继承台账 (as of 2026-08-12 ~11:0xZ — day-1 reform cycle CLOSED)

• Landed today (7 human merges, all verified on main): docs(pm-dispatch): cloud-card comms is event-driven by ruling — notes-21 verify-before-toggle, PR subscription as hard step #7794 · docs(pm-dispatch): principles-only SKILL.md rewrite — lessons distilled in place, no issue-ID citations #7936 (SKILL.md 2568→686, ID-lint, role args, adjudication protocol) · docs(agents): principles-only os-dev definition — lessons distilled in place, no issue-ID citations #7938 (os-dev.md 686→357) · [skills-path 人工合并] docs(skill): dogfood §2 — a drifted console dist now refuses to boot (#7752) #7837 · docs(pm-dispatch): subagent default for S+M with per-dispatch model tiering; seat-post read boundary #7946 (subagent backend + read boundary) · docs(agents): principles-only AGENTS.md rewrite — lessons distilled in place, no issue-ID citations #7948 (AGENTS.md 1050→818 + CLAUDE.md mirror) · docs(pm-dispatch): spec-semantic fable clause, native issue-typing + Bug-first tiebreaker, retire shift-end skill-suggestion list #7971 (spec-semantic fable clause + issue typing/Bug tiebreaker + shift-end list retirement; merged, on main at e38db3d+). All 12 member cards closed with labels stripped; both cloud sessions archived; container checkout ff'd to current main.
• In flight: none. Decision inbox: empty.
• Queue REFILLED by triage during landing (3, none dispatched): pm-dispatch: four gaps measured in one repo:cloud term — Part-of label leak, changeset-vs-published-package, PM gate lists as leads, and stale rulings #8000 (four repo:cloud-term gaps — pm-dispatch deltas), A PR body sentence saying it does NOT fix an issue still closes that issue — GitHub's keyword parser ignores negation (measured: #7888 lost this way) #7949 (GitHub Fixes-keyword negation trap), #7800 fixed the two docs a human reads; the same work-destroying recipe still ships in the os-dev agent definition, the stash hook, and the pm-dispatch SKILL #7863 (work-destroying recipe still in os-dev/stash-hook/SKILL — ⚠️ strong stale-premise candidate: those texts were rewritten today). All three cite pre-rewrite text — every one needs a stale-premise re-verification against post-docs(pm-dispatch): spec-semantic fable clause, native issue-typing + Bug-first tiebreaker, retire shift-end skill-suggestion list #7971 main before any dispatch; likely one consolidated subagent batch.
• Open findings (2): [finding] A skills-root PR was queued by a sibling session ~1h after the ADR-class ruling — a ruling recorded only on cards cannot bind sessions that haven't read it #7701, [finding] The release-blockers health metric can't tell "empty board" from "no board" — objectstack and cloud have no target:* label at all, so every round report's zero is unverified #7964 (release-blockers metric can't tell empty board from no board) — both ungraded, awaiting the triage seat.
• Next-wave plan (recorded for the next context): premise-check pm-dispatch: four gaps measured in one repo:cloud term — Part-of label leak, changeset-vs-published-package, PM gate lists as leads, and stale rulings #8000/A PR body sentence saying it does NOT fix an issue still closes that issue — GitHub's keyword parser ignores negation (measured: #7888 lost this way) #7949/#7800 fixed the two docs a human reads; the same work-destroying recipe still ships in the os-dev agent definition, the stash hook, and the pm-dispatch SKILL #7863 on current main → drop what today's rewrites already absorbed → one consolidated subagent batch (claude-fable-5, ADR-class) for the remainder. Also carried: the "ff local main before subagent dispatch" one-liner (measured gap, not yet filed as a card).

热文件串行队

.claude/skills/pm-dispatch/SKILL.md / references/** / .claude/agents/os-dev.md: queue empty, nothing in flight. Next entrant: the #8000/#7949/#7863 batch after premise re-verification.

常设承诺

1. 专题讨论 channel with the maintainer for all skills matters.
2. Skills PRs are ADR-class — AI seats draft and review only; ⛔ never merge/enqueue/arm auto-merge; the maintainer merges manually. ⚠️ The maintainer may act through shared/alternate accounts — check the seat chat before withdrawing (measured four times, 2026-08-12: note + ready-flip + enqueue via shared account = the maintainer's merge click).
3. claude-fable-5 mandatory for pm-dispatch-edit cards and spec-semantic cards (both clauses now merged SKILL text).
4. Round reports carry an "awaiting a human merge" line for every open skills PR.

说明

Founding rulings (maintainer, 2026-08-11, verbatim, untranslated):

我认为所有的 skills 任务应该单独一个项目经理负责,专题和我讨论。

所有 skills 的更新和 adr 类似,需要人工审核

所有现有 skills 相关卡片转到 新的项目经理

现在 的 项目经理skills 太长了,需要考虑怎么压缩

2026-08-12 rulings: all nine now embodied in merged skill text on main (verbatim quotes carried in the texts and on the implementing cards; this body keeps pointers, not copies).

Prior body revisions carry the founding stock-transfer table and earlier round states (edit history is the archive, per #7583).

Seat timers (current): cross-wall fallback heartbeat trig_0139XfCW1b1oY5g6T6KeHkVg (hourly cron at :50, self-bind; delete or hand over at seat exit). PR subscriptions (current session): none open (all auto-released on merge).

---

Seat created and stocked by the domain:devx seat (session session_01GZKbx4xyF7U5WXj6ch49BM), 2026-08-11.

[huangyiirene]

Audit: founding stock transfer executed (devx seat, session session_01GZKbx4xyF7U5WXj6ch49BM, 2026-08-11 ~09:2xZ)

Maintainer instruction (verbatim): 「所有现有 skills 相关卡片转到 新的项目经理」.

Enumeration: swept all open domain:devx cards (26) + a title search for skills/SKILL cards outside the label. 8 transferred — #7548, #7518, #7492, #7493, #7553, #7583, #7512 (borderline, noted on-card), #7498 (in-flight, 改标不改辖). Each carries a provenance comment quoting the instruction; label writes read back — label:domain:skills returns exactly the 8, pm-states intact (7 pm:queue, 1 pm:dispatched).

Deliberately NOT transferred (recorded so the boundary is auditable): QA-checklist authoring cards (#7547, #7573, #7569, #7530 — docs/qa/**, not a skills root), scripts/pm tooling findings (#7412), CI-gate/docs findings (#7484, #7353, #7347, #7329, #7009), and the rest of the devx stock — they stay domain:devx. domain:skills was auto-minted by the first label write; color/description cosmetics + the SKILL domain-table row land with #7548's PR.

---

Generated by Claude Code

[os-help]

Audit: seat takeover

• Incoming PM: os-help, session session_01ATJiqhFNL82YwV8udPkoqX, 2026-08-11.
• Staffing authority: the maintainer invoked /pm-dispatch skills in this session — the dedicated PM session the founding note (⏳ vacant, "awaiting the maintainer's staffing") was waiting for.
• Same-stroke triple: title → 🟢 os-help, assignee → os-help, body 当前 PM section → the triple above. Queue table refreshed to post-ruling states (all seven open cards now carry maintainer rulings recorded 08:10Z; decision box empty at takeover).
• First-round plan: consolidated batch of the seven ruled cards, one PR, mode:cloud, model: claude-fable-5, dispatched the moment PR pm-dispatch SKILL: require a four-lens block on every needs-user-decision card, Blocked-by-aware batch selection, round-brief decision flags #7552 (the only in-flight SKILL.md PR, awaiting the maintainer's manual merge) reaches terminal state. Round report goes to the maintainer in chat.

---

Generated by Claude Code