Skip to content

docs(skills): rest-channel records the advisory_report 403 re-test as transient, with the seat retry rule - #30130

Closed
objectstack-fleet[bot] wants to merge 1 commit into
mainfrom
claude/project-thread-kdvgq3
Closed

objectstack-fleet[bot] wants to merge 1 commit into
mainfrom
claude/project-thread-kdvgq3

Conversation

@objectstack-fleet

Copy link
Copy Markdown
Contributor

Requested by Irene · project thread

Fixes #29402

Clause-②: no
Metadata-protocol: none

The maintainer waived the model-tier mandate while that tier had no quota; this PR was developed below it.

维护者速读

  • 改了什么:rest-channel.md 写侧公告两行重写:L50 并入 advisory_read,L51 记 advisory_report 403 复测读数与席位规则。
  • 为什么改:卡上复测(评论 6632996128)经中继新报一次无 403,11-02 那批 403 不复现,读作瞬态,非通道常限。
  • 风险与回滚:Tier S .claude/** 文字,行数不变(82,上限 82),不进任何包,skip-changeset;回滚即 revert。
  • 先看:L50-51 两行。
  • 需要你做的:无;席位 contract review PASS 且检查全绿即落地。

What changes

.claude/skills/pm-dispatch/references/rest-channel.md, write side, two lines rewritten, none added:

  • L50 now names advisory_read beside the four advisory write operations and keeps the pointer to security-findings.md. The old L51's detail (one-time key in the seat process, ciphertext back, --reveal prints plaintext) already lives in security-findings.md L22 and lanes/security.md L19, so folding it in loses no rule. "发布归维护者" is likewise held by security-findings.md L24.
  • L51 records the reading and the seat-side rule in one line: a later advisory_report 403 was re-tested and is transient; the seat reads a 403 against the two causes security-findings.md L23 already names (the advisory write grant and the per-account daily limit on private reports), retries once on a later day, and never switches account.

security-findings.md is not edited: the reading does not make L23 wrong, it only says the 11-02 run did not hit either cause as a standing state.

Ratchets. Line count 82 → 82 (ceiling 82, the added rule line is paid by folding L51 into L50). Widths: L50 110 bytes, L51 118 bytes, under the 120-byte cap.

Why

The card's last bullet. The 11-02 runs answered HTTP 403 on every advisory_report after the first. On 2026-11-04 a fresh report through the relay succeeded with no 403, and the maintainer confirmed the lane App shows as reporter on it and on the reports of the last two days (card comment 6632996128). The seat needs a written rule for the next 403 so it neither resends into a refusal nor reaches for another account.

Verification

Head 93196356, one commit on a52761fd.

Local gates on 9319635: 17 derived, 17 run with exit 0, 0 NOT MEASURED (`dispatch-gates --ran`)
  • node scripts/pm/dispatch-gates.mjs --commands derived 17 commands for the one changed file. All ran after pnpm install --frozen-lockfile, each with its exit code captured before any pipe. check:doc-formula-expressions first exited 3 (prerequisite not built); after turbo run build --filter=@objectstack/formula --filter=@objectstack/lint it exited 0.
  • All 17 exited 0, among them check:pm-skill-ratchet, check:pm-skill-id-lint, check:skill-frame-sync, check:doc-authoring and check-comment-mask-corpus.mjs.
  • --ran printed: 17 derived famil(ies) accounted for — 17 run, 0 NOT-MEASURED (a DERIVED zero — all 17 recorded an exit code and none of them is 3).
  • node scripts/check-empty-changeset.mjs --base origin/main: exit 0. No changeset: .claude/** ships in no package, so skip-changeset.
  • node scripts/check-commit-card-trailers.mjs --range origin/main..HEAD: exit 1, found after the push. The commit message names this card and its co-author trailer carries a model name. The branch is not claude/issue-*, so AGENTS.md §3 rules out a force-push to reword it. The squash message is taken from this body, so neither line lands on main.

Human Minimum

  • 1. Schemas start as Zod; packages/spec holds no runtime logic. (Not touched.)
  • 2. Every name I added uses the casing for its kind. (No names added.)
  • 3. No formatter ran.
  • 4. One issue, one branch off an up-to-date main, one pull request.
  • 5. A change that publishes carries a changeset, and both declaration lines are filled in here and in the changeset. (Publishes nothing.)
  • 6. packages/spec touched: its generated artifacts are regenerated and committed. (Not touched.)
  • 7. No hand edits to content/docs/references/, packages/*/CHANGELOG.md or content/docs/releases/.
  • 8. Title, body and comments are in English. (The 维护者速读 section is Chinese, per the lane template.)
  • The Before Submitting list in CONTRIBUTING.md is done.

🤖 Generated with Claude Code

https://claude.ai/code/session_01J8VVAnSGyRU8EkF8acg337

…est as transient and the seat's retry rule

The live re-test on #29402 sent a fresh report through the relay with no 403,
so the earlier 403 run does not reproduce. The seat reads a later 403 against
the two causes security-findings.md already names, retries once on a later
day, and never switches account. The advisory_read line folds into the
advisory row (its key and --reveal detail already live in security-findings.md
and lanes/security.md), so the file stays at its 82-line ceiling.

Fixes #29402

Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01J8VVAnSGyRU8EkF8acg337
@objectstack-fleet objectstack-fleet Bot added the skip-changeset PR has no user-facing published change; bypasses the changeset gate label Nov 4, 2026
@github-actions github-actions Bot added size/xs documentation Improvements or additions to documentation labels Nov 4, 2026
@objectstack-fleet

Copy link
Copy Markdown
Contributor Author

domain:skills#1 · huangyiirene · session_01J8VVAnMUWA9rufXzFaEh4J

Superseded by #30132 · closed unmerged · 2026-11-04T15:22Z

This branch's only commit carried a card keyword and a co-author trailer that check-commit-card-trailers refuses. Since the branch cannot be rewritten, the same two-line change was re-cut as one clean commit on claude/project-thread-kdvgq3-r2 (PR #30132). That PR carries Fixes #29402. Nothing here is lost.


Generated by Claude Code

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

Labels

documentation Improvements or additions to documentation size/xs skip-changeset PR has no user-facing published change; bypasses the changeset gate

Projects

None yet

1 participant