Repository navigation
[governed] AGENTS.md and .claude/agents/os-dev.md disagree on who sets a PR assignee at pr_create #29367
Description
domain:spec#1 · os-steve · session_01ADHkmPi25XXKx92Xmg4viS
Filing gate: ③ escalation of the spec seat's contract review on PR #29331 (#28655), at head 5f8d7494. ⛔ Not a claim. No grade: triage sets it. Governed surfaces (Tier S/H): the fix needs a GOVERNED_APPROVERS approval.
What disagrees
.claude/agents/os-dev.mdsayspr_createcarries no assignees and the dev never sets one (the spec(object-extensionfieldPatches): refuse a field both replaced infieldsand patched infieldPatchesof one extension, key option patches by the option-value grammar, carry the whitelisted aliases, and derive the describe's member list #28655 dev followed it and declared it as a deviation).AGENTS.mdsays the PR's assignee mirrors the card's, set by the executor.
A dev following one breaks the other, and each card thread re-litigates it in its report.
Done when
One rule, stated in one place and referenced from the other: who sets the PR assignee (the dev at pr_create, or the seat at landing), and whether the relay's pr_create op accepts it.
Dedupe: none found on assignee pr_create os-dev. New.
Activity
objectstack-fleet commentedon Nov 2, 2026
triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U
Path: none
Triage: p3 ·
domain:skills·pm:queue· 2026-11-02T14:56Z⛔ Not a claim.
AGENTS.mdand.claude/agents/os-dev.mddisagree on who sets a PR assignee. Governed Tier S/H, p3.Generated by Claude Code
objectstack-fleet commentedon Nov 3, 2026
domain:skills#3 · os-support-ai · session_01WGufzhd2Ad9ANGTyAMW7pH
Claim: PM loop round 1 · 2026-11-03T03:24Z
Session:
session_01WGufzfNGqqzsW5spZyt9Xd(os-dev card thread "Align PR assignee rule across docs"; dispatched bysession_01WGufzhd2Ad9ANGTyAMW7pH)Account:
os-support-aiBranch:
claude/project-thread-hrv6lsWorktree:
objectstack-issue-29367Domain:
domain:skillsSeat:
domain:skills#3File surface:
AGENTS.md(the claim paragraph's PR-assignee parenthetical only),.claude/agents/os-dev.md(rule 2 region); no changeset (skip-changeset)Container & model:
S,mode:thread,model: CONTRACT_REVIEW_TIER — dispatch-gates --tier MANDATORY on .claude/agents/os-dev.md (clause ①, protocol semantics),effort: highClause-②: no
Metadata-protocol: none
Responsibility: n/a — not a defect card
Thread-read: 6580318036
Premise: holds
Serial constraints cleared: PR #29717 (#29709, another seat) rewrites two
AGENTS.mdroles-table rows — a different region, ordinary concurrency; no in-flight claim on.claude/agents/os-dev.mdobjectstack-fleet commentedon Nov 3, 2026
domain:skills#3 · os-support-ai · session_01WGufzfNGqqzsW5spZyt9Xd
os-dev-report
{ "issue": 29367, "status": "done", "branch": "claude/project-thread-hrv6ls", "pr": "https://github.com/objectstack-ai/objectstack/pull/29832", "head": "507f74ae724cfb4764b1a477ff9a1b0d6f8146e9", "session": "session_01WGufzfNGqqzsW5spZyt9Xd", "premise_still_valid": true, "summary": "One home for who sets a PR's assignee: .claude/agents/os-dev.md rule 2 line 38 now reads `PR ⛔ 不带 assignees(中继虽收),席位交棒才派。` (the relay's pr_create accepts the key, a dev does not send it, the seat assigns at hand-over); AGENTS.md lines 403-404 replace `the PR's assignee mirrors the card's, set by the executor at pr_create` with `nor the PR's: pr_create carries none; the seat sets it at hand-over, .claude/agents/os-dev.md rule 2`. Branch fast-forwarded onto origin/main b0ef25f5 before the single commit; merge-tree against main clean at open. Draft PR #29832, label skip-changeset, no assignees (read-back: assignees none).", "tests": "n/a — two instruction files, no code; verification = the derived gates below. Byte measure: AGENTS.md lines 403/404 = 98/113 bytes, os-dev.md line 38 = 119 bytes (cap 120). grep 'mirrors the card' AGENTS.md: no match.", "gates": "dispatch-gates --commands --repo objectstack-ai/objectstack: 23 derived (2 paths vs merge base b0ef25f5), run serially with exit codes captured before any pipe, all 23 exit 0; --ran: 23 run, 0 NOT-MEASURED. check:doc-formula-expressions first answered exit 3 (formula/lint dist not built, nothing measured); after turbo build --filter=@objectstack/formula --filter=@objectstack/lint it ran, exit 0. check:pm-skill-ratchet: AGENTS.md 1050 lines (ceiling 1050, headroom 0), os-dev.md 405 lines (ceiling 405, headroom 0), every line within 120 bytes. check:pm-skill-id-lint: 109 files clean. CI check-runs: pending at report time.", "line_budget": "AGENTS.md net 0 (1050 → 1050), os-dev.md net 0 (405 → 405). The order allowed os-dev.md ≤ +1 and read AGENTS.md's ceiling as 1117; check-skill-line-ratchet.mjs holds 405 and 1050 at headroom 0, so both edits were made in place with no re-wrap of untouched lines.", "dogfood_battery": "n/a — no `!` in the PR title and no (narrowing) arm", "mcp_calls": "3 — GitHub MCP reads only: issue_read get (card body), issue_read get_comments x2 (triage + Claim). No MCP write tool.", "api_writes": "2 relay writes as objectstack-fleet[bot]: pr_create (#29832, draft, labels [skip-changeset], no assignees, 6394 bytes read back identical), comment (this report on #29367 via post-stamped). Plus 1 git push of the pinned branch. No pr_request_reviewers.", "writes": "git push x1; relay pr_create x1; relay comment x1", "files_changed": [".claude/agents/os-dev.md (line 38, net 0)", "AGENTS.md (lines 403-404, net 0)"], "deviations": [ "os-dev.md edited at net 0 instead of the order's allowed +1: the ratchet ceiling (405) sits at the current count, so the +1 would have gone red; the relay-accepts clause was fitted into line 38 within 120 bytes.", "The harness session's assign-and-request-review step was skipped on purpose (os-dev rule 2, the rule this PR makes the one home).", "Worktree ../objectstack-issue-29367 kept for a patch round." ], "open_questions": [], "security_findings": "none", "out_of_scope_findings": [] }objectstack-fleet commentedon Nov 3, 2026
domain:skills#3 · os-support-ai · session_01WGufzhd2Ad9ANGTyAMW7pH
ACCEPT → PR #29832 comment 6597290481 · Tier H，已请 os-zhuang / hotlong 审批，获批后席位落地。