Repository navigation
pm-dispatch SKILL Model tiering: mandatory claude-fable-5 for spec SEMANTIC cards — accept/reject behavior or public-surface change (maintainer ruling 2026-08-12) #7945
Description
Filed by the skills seat (session session_012Gg6rMAti8ZaueWb6BRsDn), 2026-08-12, executing a maintainer ruling from the seat 专题 chat.
Provenance (maintainer, verbatim, untranslated)
本项目是协议驱动的,是否建议spec 协议相关的开发都是用 Fable 模型,我发现实际派发的时候很多都是 opus
同意,就按语义面收窄,立卡并通知 spec 席
The approved shape (recommended by this seat, accepted as-is): fable follows the criterion, not the domain:spec label.
The clause to add (Model tiering section)
- Mandatory
claude-fable-5for any card whose change alters the contract's accept/reject behavior or grows its public surface — thedomain:specsemantic lane (schema shapes,contracts/**, strictness, the behavior half of retirements). Same no-downward-discretion form as the existing pm-dispatch-edit clause. Rationale on record:packages/specis the single contract; an error there poisons every compile surface, downstream repo, AI-authored app and stored metadata — the most expensive error class in the repo, and the layer that makes AI-written apps structurally hard to get wrong. - Explicitly unchanged:
domain:spec-surface(text-only, accept-set byte-identical — machine-checked) stays sonnet/opus;domain:spec-toolingstays opus default, upgrading to fable when the deliverable is a gate's semantics design; downstream-adaptation cards (consumer-side sweeps of an already-fixed contract) stay opus default, upgrade-on-doubt as today. - The criterion is the existing mechanical test already in the SKILL (the spec/spec-surface split; the adjudication boundary test) — cite it, do not restate a second copy.
Landing notes
- Blocked-by: pm-dispatch SKILL: M-grade cards default to the subagent backend with per-dispatch model tiering — cloud reserved for L/XL, long-lived, browser-verify (maintainer ruling 2026-08-12) #7944
- Same-file serial: SKILL.md hot-file queue position 2, behind the in-flight pm-dispatch SKILL: M-grade cards default to the subagent backend with per-dispatch model tiering — cloud reserved for L/XL, long-lived, browser-verify (maintainer ruling 2026-08-12) #7944 batch PR₁ (which is editing the same Model-tiering/backend region — dispatching earlier guarantees rework).
⚠️ Ratchet headroom 0 — offset in-PR. - Effective immediately ahead of the text landing, via cross-seat notice on the spec semantic seat post ([PM seat] domain:spec — 🟢 os-tesla · session_013odhA26dvzy7pQmevdTTvy #6017) with the ruling verbatim.
- Dispatch
model: claude-fable-5(mandatory clause — this card edits.claude/skills/pm-dispatch/**). Size S — subagent lane per the pm-dispatch SKILL: M-grade cards default to the subagent backend with per-dispatch model tiering — cloud reserved for L/XL, long-lived, browser-verify (maintainer ruling 2026-08-12) #7944 revision. - ADR-class: draft PR, human merge.
Activity
os-zhuang commentedon Aug 12, 2026
Batch conversion — this card becomes the carrier of the closing consolidated dispatch (maintainer prompt in seat chat, 2026-08-12, verbatim: 「剩余的 skills 是否应该一个批次处理完。」; sequencing is the PM's call).
Composition: ONE os-dev subagent (
mode:subagentper the #7944 revision,model: claude-fable-5mandatory), ONE PR —Fixes #7945,Fixes #7947— landing both ruled S-grade SKILL.md deltas (this card's spec-semantic fable clause + #7947's issue-typing/Bug-tiebreaker) with a per-member checklist and ratchet-0 in-section offsets.Dispatch trigger: PR #7946 (the in-flight batch PR₁) reaching MERGED — both deltas edit the sections it rewrites.
Blocked-by: #7944stands until then; #7947 is hereby a member, not an independent candidate.Generated by Claude Code
os-zhuang commentedon Aug 12, 2026
Batch composition update (maintainer approval 2026-08-12, verbatim: 「同意」): #7950 (retire the mandatory shift-end SKILL-suggestion list; three report classes, default zero) joins as the third member. The closing batch is now: one subagent (
model: claude-fable-5), ONE PR —Fixes #7945,Fixes #7947,Fixes #7950— per-member checklist, ratchet-0 offsets (note #7950 is net-negative). Dispatch trigger unchanged: PR #7946 reaching MERGED.Generated by Claude Code
os-zhuang commentedon Aug 12, 2026
Claim: PM loop round 5 (skills seat) — closing consolidated batch carrier
Session:
session_012Gg6rMAti8ZaueWb6BRsDnBranch:
claude/issue-7945-pm-skill-closing-batch(one PR:Fixes #7945 / #7947 / #7950)Worktree: dedicated worktree inside the PM session's container (subagent backend, per the now-landed revision)
Domain:
domain:skillsFile surface:
.claude/skills/pm-dispatch/SKILL.md(Model tiering + triage duty + selection sections) +references/seat-post-protocol.md(seat-exit checklist). Stop on breach; explain in the report.Container & model: S docs batch,
mode:subagent,model: claude-fable-5(mandatory clause — no downward discretion)Serial constraints cleared: predecessor PR #7946 MERGED (merge commit
6ff179dverified on origin/main; SKILL.md = 686 lines, backend revision live); PR #7948 next in the merge queue touches AGENTS.md/CLAUDE.md only — no file overlap with this batch; zero other open PRs touch the pm-dispatch tree (checked at claim time). Members #7947 / #7950 absorbed per the composition comments; #7945'sBlocked-by: #7944cleared by the merge.Landing: ADR-class — draft PR, human merge.
Generated by Claude Code
os-zhuang commentedon Aug 12, 2026
Dispatched (PM loop round 5, skills seat): os-dev subagent,
model: claude-fable-5, launched 2026-08-12 ~08:4xZ. One draft PR expected (Fixes #7945 / #7947 / #7950), branchclaude/issue-7945-pm-skill-closing-batchoff6ff179d+. All three deltas fully ruled; #7950's retirement funds the line budget at ratchet headroom 0. Collection: os-dev-report comment here + direct return channel; ADR-class terminal (human merge).Generated by Claude Code
os-zhuang commentedon Aug 12, 2026
{ "issue": 7945, "status": "done", "branch": "claude/issue-7945-pm-skill-closing-batch", "pr": "https://github.com/objectstack-ai/objectstack/pull/7971", "premise_still_valid": true, "summary": "Closing consolidated batch landed as ONE draft PR (Fixes #7945 / #7947 / #7950, each on its own line). #7945: Model tiering now carries mandatory clause 2 - claude-fable-5 for any card altering the contract's accept/reject behavior or growing its public surface (domain:spec semantic lane), citing the existing adjudication boundary test and the one-package-three-seats split with no second copy; spec-surface/spec-tooling/downstream-adaptation tiers explicitly unchanged. #7947: triage seat is the single authoritative producer of native issue types Bug/Feature/Task with the mechanical criterion, Bug-without-repro stays out of pm:queue, no mass backfill, three named consumers; Bug-first tiebreaker (not a ladder) added to executor selection within ordinary queue items. #7950: exit-checklist item 7 retired in references/seat-post-protocol.md, shift-end reporting narrowed to three classes default zero, lessons-learned prose inadmissible, plus the corresponding principle line in SKILL.md's seat-post section. Ratchet held at exactly 686/686 with all offsets inside touched sections (compressed prose whose detail lives verbatim in references/ or the os-dev frontmatter comment - each verified before cutting); ceiling untouched. Branch merged past PR #7948's landing (no file overlap); gates re-run green on the merge result. ADR-class: draft PR left for human merge - no queue, no auto-merge. skip-changeset applied via read-then-union (labels read back: size/m, skip-changeset).", "tests": "All five named gates run locally under the shared verify lock, each with --self-test first: check:pm-skill-ratchet GREEN 'SKILL.md is 686 lines (ceiling 686; headroom 0)' (self-test 5/5); check:pm-skill-id-lint GREEN '9 file(s) clean (pattern /#[0-9]{3,}/g)' (self-test 9/9); check:skill-frame-sync GREEN '4 copies of the decision frame are structurally isomorphic across 3 files' (self-test 12/12); check:doc-authoring GREEN '375 files clean - no bare metadata literals'; check:nul-bytes GREEN 'scanned 7283 text file(s) ... no raw ASCII control bytes' plus targeted control-char grep of both edited files: clean. CI on the PR: in_progress at report time (honest reading; PM owns convergence).", "open_questions": [], "out_of_scope_findings": [] }Generated by Claude Code
os-zhuang commentedon Aug 12, 2026
Review: ACCEPT — PR #7971 (closing batch), ADR-class terminal: awaiting a human merge
Verified against the branch and GitHub, not the report:
main(post-docs(agents): principles-only AGENTS.md rewrite — lessons distilled in place, no issue-ID citations #7948) ✓, first linesFixes #7945 / #7947 / #7950each on its own line ✓; 2 files (SKILL.md +references/seat-post-protocol.md), ±86/82.#[0-9]{3,}= 0. All three deltas located: spec-semantic fable clause (mandatory clause ②, citing the existing boundary test — no second copy), triage issue-typing paragraph (single producer, mechanical criterion, Bug-repro discipline, no-backfill, three named consumers) + Bug-first tiebreaker in selection, and the exit-checklist item-⑦ retirement with the three-classes-default-zero rule in the reference file plus its principle line in SKILL.md.references/or the os-dev frontmatter (the PR body lists each cut with its verified destination); dropped-with-rationale list is empty — nothing from the three cards was dropped, and the ruling quotes are carried verbatim.skip-changesetsurvived, read back); ESLint + TypeScript in_progress at review time — a red triggers a patch round before merge; this seat is subscribed.Terminal state: the PR stays visibly hanging — the maintainer merges by hand. On landing: #7945 / #7947 / #7950 close via Fixes; the skills lane's dispatch queue reaches zero.
Generated by Claude Code