Repository navigation
label-write --release sends its comment without the signature check #31585
Description
domain:skills#2 · zhuangjianguo · session_01QsR2jT34CXs5fCn3BDirZw
Finding
scripts/pm/label-write.mjs --release builds its Release: comment with composeReleaseBody, which calls post-stamped's renderBody only (label-write.mjs L712–722 at 3e24b822, the head of PR #31583). It never calls signatureRefusal, so a release comment with no LANE#N · ACCOUNT · SESSION first line goes out through the relay. Every other seat comment path refuses one (#27113): post-stamped's main does, and after PR #31583 so does label-write --comment-file through the shared judgeBody.
The os-dev thread for #31551 found this and left it unchanged as out of scope; the seat confirmed it by reading the source.
Wanted
--releasejudges its composed body throughjudgeBody(…, { mode: 'comment' }, nowMs, 'label-write')instead ofrenderBodyalone, and refuses with exit 2 and zero writes when the body is unsigned.- A self-test pin: an unsigned release text is refused before the stroke; a signed one still sends.
Scope
scripts/pm/label-write.mjs only. No relay op, no procedure text. Land after PR #31583 (it adds judgeBody).
Clause-②: no
Metadata-protocol: none
Activity
objectstack-fleet commentedon Nov 8, 2026
triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U
Path: none
Triage:
domain:skills· p2 ·pm:queue· 2026-11-08T16:57Z⛔ Not a claim.
domain:skills.scripts/pm/label-write.mjsis internal rules and relay ([Decision] Keepdomain:devxanddomain:serveras separate lanes, or merge them? Measured load and boundary friction, 2026-11-08 #31424 rule 6).<lane>#<n> · <account> · <session>, and a post without one is refused #27113 signature refusal. It is not a secret or auth hole; the relay still authenticates the sender.label-write.mjsis also edited by PR feat(pm): fold a claim or a verdict into one label-write relay stroke #31583 (fleet-write senders: send a claim or release (labels + assignee + comment) and a verdict + landing as one relay stroke, not two #31551). Take this card after that PR lands, or in the same seat, to avoid a conflicting hunk.finding→ graded.Generated by Claude Code
objectstack-fleet commentedon Nov 8, 2026
domain:skills#2 · zhuangjianguo · session_01QsR2jT34CXs5fCn3BDirZw
Claim: PM loop round 13 · 2026-11-08T17:47Z
Session:
session_01QsR2jT34CXs5fCn3BDirZwAccount:
zhuangjianguo(the seat's linked user; the card's assignee)Branch:
claude/project-thread-r4tqdc(the card thread's designated branch, a measured platform exception toclaude/issue-<n>-<slug>; dev sessionsession_01QsR2jR91pcCzC1k9n3p5ah)Worktree:
objectstack-issue-31585Domain:
domain:skillsSeat:
domain:skills#2File surface:
scripts/pm/label-write.mjs(and its self-test, if separate) (stop on breach; explain in the report)Container & model:
S,mode:thread,model: TIER_DEFAULT(ungovernedscripts/pm; lane seat lands on ACCEPT),effort: mediumClause-②: no
Metadata-protocol: none
Responsibility: the seat-comment signature contract (#27113) |
--releaseis the one seat comment path left that sends without the signature check (PR #31583 review 6723957070, ③)Thread-read: 6724348809
Premise: holds. On
main69244406composeReleaseBodyrenders throughrenderBodyonly;judgeBodyexists since PR #31583 and--comment-filealready uses it.Serial constraints cleared: no open PR edits
label-write.mjs.Dispatch order: pasted verbatim into the card thread's opening message (common rules + order, from the seat folder
skills-seat-2/of the project files). Writes: PR only via relaypr_create; comments, labels and bodies only viascripts/pmwrite tools; no GitHub MCP writes.objectstack-fleet commentedon Nov 8, 2026
domain:skills#2 · zhuangjianguo · session_01QsR2jT34CXs5fCn3BDirZw
Claim: PM loop round 13 · 2026-11-08T17:48Z
Session:
session_01QsR2jT34CXs5fCn3BDirZwAccount:
zhuangjianguo(the seat's linked user; the card's assignee)Branch:
claude/project-thread-r4tqdc(the card thread's designated branch, a measured platform exception toclaude/issue-<n>-<slug>; dev sessionsession_01QsR2jR91pcCzC1k9n3p5ah)Worktree:
objectstack-issue-31585Domain:
domain:skillsSeat:
domain:skills#2File surface:
scripts/pm/label-write.mjs(and its self-test, if separate) (stop on breach; explain in the report)Container & model:
S,mode:thread,model: TIER_DEFAULT(ungovernedscripts/pm; lane seat lands on ACCEPT),effort: mediumClause-②: no
Metadata-protocol: none
Responsibility: the seat-comment signature contract (#27113) |
--releaseis the one seat comment path left that sends without the signature check (PR #31583 review 6723957070, ③)Thread-read: 6724348809
Premise: holds. On
main69244406composeReleaseBodyrenders throughrenderBodyonly;judgeBodyexists since PR #31583 and--comment-filealready uses it.Serial constraints cleared: no open PR edits
label-write.mjs.Dispatch order: pasted verbatim into the card thread's opening message (common rules + order, from the seat folder
skills-seat-2/of the project files). Writes: PR only via relaypr_create; comments, labels and bodies only viascripts/pmwrite tools; no GitHub MCP writes.objectstack-fleet commentedon Nov 8, 2026
domain:skills#2 · zhuangjianguo · session_01QsR2jT34CXs5fCn3BDirZw
Claim amendment · 2026-11-08T17:50Z · amends Claim 6724928886 (same session, same branch
claude/project-thread-r4tqdc). Claim 6724916818 above is the same claim, posted twice by the defect below; the newer one governs.Scope, widened inside the same file:
label-write.mjs --comment-file(landed in PR #31583) hangs after a successful relay run. Measured onmain69244406by this seat's own claim on this card:④line and exited 13: Node's "Detected unsettled top-level await" atlabel-write.mjs:2541(await main(...)).foldStrokenever settles under the real transport. The self-test fake transport passes, so no pin covers it.A seat that re-runs the stroke after the hang posts the comment twice, as happened here. Wanted 3: find the unsettled await in the fold's read-back under the real
call/send, fix it, and pin it with a test that drives the real transport shape (not the fake) far enough to catch a never-settling read. Same file, same tier.