Skip to content

[finding] The generic seat-post protocol and the SKILL.md seat mutex still describe a single-occupancy seat, which the director ruling retires for that seat #16343

Description

@os-steve

Filed by the dev executing the director-charter ruling on #16172 as an out-of-scope finding: the ruling was executed in .claude/skills/pm-dispatch/references/lanes/director.md only (the one file that flight held). Every sentence below lives in a file another flight holds, so none of them was touched. ⛔ No domain:* and no assignee — the skills seat triages this.

What the ruling says

The director seat is a decision function: any number of summons run concurrently, no closing brief is owed before another summon starts, and a second summon neither waits for, stands down nor supersedes another. The takeover (接管) and incumbent (在任者) notions are retired for that seat.

The generic protocol text that still assumes exactly one occupant

.claude/skills/pm-dispatch/references/seat-post-protocol.md — the file the director charter points at with 「四段照旧」. Its status vocabulary has no word for a seat with N summons open, which is the specific gap the executing brief asked to be reported:

  3:见 SKILL.md 〈座位贴协议〉;本文是接管、移交、回收那一刻要查的协议细则。
  7:- 四段的现值由在任座位在写侧刷新时收缩;⛔ 岗位说明 PR 不代改贴正文。
  9:- 状态词表:`🟢 <github-login>` 在任、`🟢 Routine`、`⏳ vacant`、`⏸️ paused`。
 11:- assignee = 在任 PM 的 GitHub 账号,退场或回收时摘除;无 assignee = 空缺,与标题互为校验。
 13:- 当前 PM 段登记三元:GitHub 账号(`get_me` 自查)、会话 ID 或 Routine ID、上任时刻。
 14:- 审计与认领评论的作者字段是平台盖章的硬读数,接管仲裁与活性判定优先用它对账正文。
 16:- 接管压缩:接任 PM 把交接的有效内容按四段模板重述为新正文,编辑历史即存档。
 34:- 收班序列:贴正文与标题刷新在先,简报评论是最终一笔;继任的释放判据 = 简报在。
 38:- 标题词表不扩,收班座位恒 `⏳ vacant`;留守登记(谁、哪些尾巴)= 说明段一行。

Line 9 is the load-bearing one: the four words are 🟢 <github-login> / 🟢 Routine / ⏳ vacant / ⏸️ paused, and line 38 forbids extending the vocabulary. A director seat post with three summons open can be written under none of them, and line 11's assignee-title cross-check reads such a post as a half state. The charter line landed for #16172 says the status block carries one line per summon with no occupant, which the generic protocol cannot currently express.

The mutex and takeover sentences in the files other flights hold

.claude/skills/pm-dispatch/SKILL.md:

 78:- 开轮互斥:fire 开始、首个写动作之前读四个读数,未收班活动任一未满一轮 ⇒ 自退。
 87:- 收班简报是前任不再写的显式声明,是释放标记不是锁:简报即最新事件 ⇒ 立即坐席。
 88:- 维护者明示召唤是仲裁:有简报径直坐席;无简报才走保守确认,确认终止即坐席。
272:- 接管/移交 = 改正文 + 审计评论(评论只作审计不承载状态),接管压缩是标准步骤。
280:- ⛔ 永不从贴的 `updated_at` 或标题推死活;活性只认在任者自己的产出。

.claude/skills/pm-dispatch/references/core-rules.md:

 19:- 开轮互斥读四个读数:收班简报、开轮标记、本车道最新他会话认领、最新关闭卡的认领。

These stay TRUE for domain:* seats, which the ruling did not touch — line 88 is the generic arbitration rule and the ruling is explicitly about the director seat alone. What is open is whether each of them needs a director carve-out, since a director summon that runs line 78 today finds an active session with no closing brief and self-retires, which is exactly what the ruling says it must not do.

A stale quotation left by this edit

scripts/pm/check-skill-line-ratchet.mjs:720 quotes the director charter's authorization sentence verbatim inside the CEILINGS comment:

720:  // file: 「项目总监 由人工定期指挥,负责处理 contract-review 类别的和需要决裁的。」) — a

The charter's line was reworded in the same stroke that folded the summon guard into it, so the quotation no longer matches the file it quotes. It is a comment, not a pin, and the ratchet is green either way; the file is held by another flight, so it is reported rather than fixed.

Dedupe

Listed all open issues over the REST list endpoint (3 pages, 300 entries) and grepped titles for seat / summon / mutex / occupancy / 座位 / director / handover / charter. The neighbours are #16172 (this ruling's own card), #16004 (the mutex's reading 3 misses a session re-engaged on a pm:queue card — a different reading, same mutex) and #16197; none of them is this. Control: the same grep returns live hits, so the empty result is a reading rather than a broken query.


Generated by Claude Code

Activity

  1. claude commented on Sep 6, 2026

    @claude
    Contributor

    Blocked-by: #16004

    Graded by the skills seat (15:1xZ): domain:skills, priority:p2, documentation + finding kept; footer restored through REST (the MCP create path drops it). Ruling recorded at flight E's ACCEPT on PR #16346: option A — one carve-out sentence in SKILL.md 〈入口与角色〉 naming the director seat exempt from readings one and two and the 自退 rule, the generic mutex untouched for domain:* seats; seat-post-protocol.md gains a status-word form for a multi-summon seat (or the director line names its own), and the ratchet map's comment quoting the old authorization sentence is refreshed. Held behind flight A's PR #16339 (SKILL.md, core-rules.md) and flight B (the ratchet map); dispatched as one flight after both merge.


    Generated by Claude Code

  2. claude commented on Sep 7, 2026

    @claude
    Contributor

    Claim: PM loop round 9 — flight Z, the SKILL.md batch: five folds under the unchanged line pins (SKILL.md 811, core-rules 150, and each named reference's own ceiling), one governed PR. (1) #16386 乙 — the human-floor row gains the one-file exception for references/platform-readings.md: an increment raises the ceiling to the landed line count without a decision card provided the seat's acceptance comment shows every fact verified and deduplicated (candidates / landed / already-present / declined), one line per fact, no re-wrapping counted; recorded in ruledRaises citing the standing ruling (5563909225). (2) #15803 §A option (c) — batch independence reads FILE SURFACE, not package; the exemption clause becomes the rule (region-level file-surface declaration in the claim; merge main before opening the PR and again after a sibling lands; conflicts arbitrated by the merge queue, no manual PM ordering; same-file strictly serial), with the sentence that cross-file semantic coupling is caught by the queue's CI (5563910163). (3) #16352 — the selection order reads priority:p1/p2/p3 where it reads p0 today, in the order the card's ruling states. (4) #16372 — re-read the PR head before trusting any check_suite.completed event. (5) #16343 — the director carve-out in the seat mutex, the seat-post protocol's multi-summon status form, the ratchet-map comment refresh.
    Session: session_019RfFHiRCSs3JXLK4cwcfox
    Branch: claude/issue-16386-skill-batch-62
    Worktree: objectstack-issue-16386
    Domain: domain:skills
    File surface: .claude/skills/pm-dispatch/SKILL.md, references/core-rules.md, and whichever of references/seat-post-protocol.md and the ratchet map comment in scripts/pm/check-skill-line-ratchet.mjs the five cards name — measured per card before editing; folds only, no ceiling raise.
    Container & model: L, mode:subagent, fable — dispatch-gates.mjs --tier: MANDATORY on .claude/skills/pm-dispatch/SKILL.md; on a fable 429 the seat's exemption applies, opus, recorded here.
    Clause-②: no
    Serial: devx's PR #16271 (draft, mergeable_state: dirty) collides on 候选与批次 and owes the re-fold per note 5560027553; this flight branches from origin/main and takes no proof against a dirty head — whichever merges second re-folds. T's PR #16437 is in the queue on the ratchet script's self-test region; merge origin/main in after it lands or prove clean twice. Governed ⇒ draft PR, in-seat contract-tier review, then the seat requests os-zhuang + hotlong, adds needs-user-decision and posts the final 维护者速读; human merge. The PR carries Fixes #16386, Fixes #15803, Fixes #16352, Fixes #16372, Fixes #16343.


    Generated by Claude Code

  3. claude commented on Sep 7, 2026

    @claude
    Contributor

    Report pointer (flight Z): this card is one of the five folded into draft PR #16486 (head 9390702); the os-dev-report for the flight is the anchor comment on #16386 — #16386 (comment) — and the PR body lists this card's old line, new line and the fold that paid for it.


    Generated by Claude Code

  4. github-actions commented on Sep 7, 2026

    @github-actions
    Contributor

    os-closed-card-sweep — machine-findable marker for this generated comment.

    Removed the pm-loop state label(s) this closed card no longer claims: pm:dispatched.

    A state label claims work is in flight. This card is closed on a merged delivery, so the claim
    is stale; every other label is left exactly as it was found. Nothing here is a judgement about
    the card, and no verdict-bearing label is ever touched by this sweep.

    posted by half-state-patrol run 34097314575 · trigger schedule

    Generated by Claude Code

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions