Skip to content

docs(pm-dispatch): resolve the self-contradiction left by #6959 and #6886 landing one merge apart (#5469) - #6974

Merged
os-zhuang merged 1 commit into
mainfrom
claude/lane-task-statistics-g74nn1
Aug 9, 2026
Merged

docs(pm-dispatch): resolve the self-contradiction left by #6959 and #6886 landing one merge apart (#5469)#6974
os-zhuang merged 1 commit into
mainfrom
claude/lane-task-statistics-g74nn1

Conversation

@os-zhuang

Copy link
Copy Markdown
Contributor

Part of #5469. Repairs a defect currently on main that neither PR could have caught alone.

What is wrong on main right now

#6959 (0b7093e) and #6886 (64fa424) landed one merge apart into the same file. Both are individually correct; the merged tree is not. Git text-merged them with exit 0 and zero conflict markers, and the result gives triage two opposite instructions about the same three paths:

Where What it says today
Package-family table, spec-tooling row (#6959) The lane has a row; packages/lint / scripts/ / content/docs/** route by the "orbits the spec contract" criterion
Roster table row (#6886) ⚠️ **故意无行** — deliberately no row
### domain:spec-tooling section (#6886) Title says 包家族表故意不给它行(#5469 未闭); the three paths are ⚠️ 未裁; cards landing there must FLAG 回 #5469,⛔ 不自行二选一

#6886's text was written before the 2026-08-09 ruling and is stale, not wrong-headed — that session's own recommendation on #5469 was option A, which is what the maintainer ruled. Only the prose needs to catch up.

This is the same silent-swallow shape the file already documents for merge=os-regen (recorded on #6018): the merge succeeds, nothing is flagged, and the damage is only visible by reading the merged tree. Here it swallowed semantics rather than bytes, so no regen driver would have caught it.

What changed

.claude/skills/pm-dispatch/SKILL.md, docs-only, +17/−8:

  1. Roster row✅ 有行(2026-08-09 裁决后补,见下一段).
  2. Section title + new lede — retitled to 三处争议面已按判据裁定, with an explicit reading instruction: this section is history, not a to-do; the live routing rule is the package-family table. That framing is what stops the next drift, since the section's value is precisely the archaeology of two superseded rulings.
  3. Disposition rewritten — the 2026-08-06 ruling and the three facts that overtook it are kept verbatim (that record is the section's point). What follows now registers the 2026-08-09 criterion ruling instead of declaring the question open, and notes it extends the 2026-08-07 split's own criterion rather than inventing a second rule.
  4. FLAG exit narrowedFLAG 回 #5469 (a closed question) → rule 4's ordinary mislabel path for genuinely unclear cards. ⛔ 不由 dev 代拍 is retained.

Not touched: the roster's retired-labels table, the label-creation setup block, and the two-path uncontested routing — all still accurate.

Gates

check:skill-frame-sync ✓ (4 copies isomorphic, 3 axes), check:skill-compatibility-version ✓. Post-edit grep for 故意无行 / 未裁 / FLAG 回 #5469 returns only the new history sentence. The published mirror skills/objectstack-pm-dispatch/SKILL.md contains no spec-tooling text, so no sync obligation. Releases nothing → skip-changeset.

Branch restarted from origin/main after #6959 merged, per the merged-PR follow-up rule.


Generated by Claude Code

…-09 裁决 (#5469)

两个 PR 相隔一次合并落在同一文件:#6959 给 domain:spec-tooling 补了包家族表行
(维护者 2026-08-09 判据切分裁决),#6886 同批落地的花名册与沿革段则写于裁决之前,
说该行「故意无行」、三处争议面「未裁」、落在那里的卡要「FLAG 回 #5469」。文本合并
exit 0、零冲突标记,合并后的树因此对分诊同时下两道相反指令 —— 正是本文件
merge=os-regen 那条记的静默吞并形状,只不过这次吞的是语义不是字节。

修:花名册行改 ✅ 有行;沿革段标题与首段点明「本节是沿革不是待办,生效规则在包家族
表」;2026-08-06 旧裁决保留(它被事实推翻的记录仍有价值),其后的处置改为登记
2026-08-09 判据切分裁决,FLAG 出口从 #5469 收回到 rule 4 常规误标路径。

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BPWqbmEFU8gJepBJTHESXd
@vercel

vercel Bot commented Aug 9, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment
Project Deployment Actions Updated (UTC)
objectstack Ignored Ignored Aug 9, 2026 6:57am

Request Review

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/s skip-changeset PR has no user-facing published change; bypasses the changeset gate

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants