Repository navigation
pm-dispatch: three more PR-body rewrites and the rest-channel landing rule do not name the issue_patch body door #31050
Description
domain:skills#3 · huangyiirene · session_01J8VVAj7CvjkGXvgXzd6RZ7
Filing gate: out-of-scope should-fixes S2–S4 of the contract review of PR #31049 (card #29285). They fall outside that PR's claimed files. ⛔ Not a claim. Triage sets the grade and the lane.
class: c · reach: every seat that rewrites a pull request body, and every reader of the handover gate's header.
What is left after PR #31049
PR #31049 teaches the PR-body door (a relay issue_patch through scripts/pm/fleet-write/dispatch.mjs, signed by the trailer on the last non-empty line) in .claude/skills/pm-dispatch/references/seat-post-protocol.md and the claim template's Metadata-protocol: row in SKILL.md. Its contract review found four places outside those two files that still disagree or stay silent:
references/rest-channel.mdL16 says 「直调dispatch.mjs仅落地」 (a directdispatch.mjscall is for landing only).SKILL.mdL87 places landing after ACCEPT, but the PR-body rewrite happens before ACCEPT (owed — FILES→#N, a round-2 body update, a flake note). The rule needs an explicit exception for the single-op PR-bodyissue_patch, or a pointer toseat-post-protocol.md's PR-body line.scripts/pm/check-handover.mjsheader (≈L33) still namesseat-post-protocol.md收班 as the home of the sign-off hand-back rule that narrows (a)'s stay-behind. PR docs(pm-dispatch): teach the seat the PR-body issue_patch door #31049 deletes that line as a restatement; its surviving home isreferences/seat-lifecycle.mdL44 (and the handover checklist step 2 / L38 for the default hand-back). Repoint the header.references/platform-readings.mdL290 (the seat edits a PR body to record a flake reason, and itseditedrerunsCheck Changeset) and.claude/agents/os-dev.mdL66 (席位代写 of a report-named body change) are PR-body rewrites that name no door. Point each atseat-post-protocol.md's PR-body line.- No-footer rule reach (delta review of PR docs(pm-dispatch): teach the seat the PR-body issue_patch door #31049, D2).
seat-post-protocol.md's PR-body line names the trailer and the door but not "send no footer" (platform-readings.mdL336–341). OnlySKILL.mdL313 points to that reading. A rewrite reached through item 3's pointers would miss it, so the pointer (or the PR-body line) should carry it. While there, reconcileplatform-readings.mdL341 ("a 58 B bare footer" appended on a PR-bodyPATCH) with the live reading in pm relay: no door can rewrite a PR body after round 1; the dev writes it once (os-dev.md), and a seat issue_patch must open with its signature, which breaks the Fixes-first body shape #29145 and PR docs(pm-dispatch): teach the seat the PR-body issue_patch door #31049's own body, which ended with the trailer and no appended footer.
Level: internal agent tooling (.claude/** and scripts/pm/** comment only), skip-changeset, Clause-②: no, Metadata-protocol: none.
Activity
objectstack-fleet commentedon Nov 6, 2026
triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U
Path: none
Triage: audit · p2 ·
domain:skills· 2026-11-06T21:54Z⛔ Not a claim.
issue_patchdoor that PR docs(pm-dispatch): teach the seat the PR-body issue_patch door #31049 teaches. The places:rest-channel.mdL16 ("landing only"), thecheck-handover.mjsheader pointer,platform-readings.mdL290 andos-dev.mdL66 (rewrites that name no door), and the no-footer rule's reach.domain:skills(.claude/skills/pm-dispatch/**; the script header is a comment).findingremoved,pm:queue.Generated by Claude Code
objectstack-fleet commentedon Nov 6, 2026
domain:skills#3 · huangyiirene · session_01J8VVAj7CvjkGXvgXzd6RZ7
Claim: PM loop round 8 · 2026-11-06T22:49Z
Session:
session_01J8VVAj7CvjkGXvgXzd6RZ7Account:
huangyiireneBranch:
claude/project-thread-0ili91(the card thread's designated branch; dev sessionsession_01J8VVAkpKySQ5xGrSjjjnVw)Worktree:
objectstack-issue-31050Domain:
domain:skillsSeat:
domain:skills#3File surface:
.claude/skills/pm-dispatch/references/rest-channel.md(L16);.claude/skills/pm-dispatch/references/platform-readings.md(L290; L336–341 only as the card's item 4 needs);.claude/agents/os-dev.md(L66);.claude/skills/pm-dispatch/references/seat-post-protocol.md(L31);scripts/pm/check-handover.mjs(header comment ≈L33 only, no code); deletion-only restatement edits elsewhere in those files to pay the line ratchet, each with its surviving home named in the PR body (stop on breach)Container & model:
S,mode:thread, project-default model,effort: medium.dispatch-gates --tieris run by the thread and reported; any mandate above this thread's is waived by the maintainer's standing order in this project (2026-11-03) and said so in the PR body; this seat runs the contract review.Clause-②: no
Metadata-protocol: none
Responsibility:
n/a — residue of PR #31049's teaching (its contract review S2–S4, D2); objectstack carries the fixThread-read: 6696823327
Premise: holds on
mainb5106af62: rest-channel.md L16 reads 「直调dispatch.mjs仅落地」; platform-readings.md L290 and os-dev.md L66 rewrite a PR body without naming the door; check-handover.mjs's header still names seat-post-protocol 收班 as the hand-back rule's home.Serial constraints: none.
No changeset (nothing published):
skip-changeset.Fixes #31050. Tier read bycheck-governed-mergesat the PR.objectstack-fleet commentedon Nov 6, 2026
domain:skills#3 · huangyiirene · session_01J8VVAj7CvjkGXvgXzd6RZ7
ACCEPT → PR #31058 comment 6698729563 (seat check and maintainer summary). Tier S: same-head PASS on record after one FAIL round, CI all green; the seat flipped it ready and armed auto-merge. Residue filed as #31059.