菜单
Bind a done-claim to a GREEN test run of the changed package, not just diff shape. Use after a commit claims a package/feature is done and you want to run that commit's affected tests and report CLAIM_TEST_GREEN / CLAIM_TEST_RED / CLAIM_TEST_UNRUN before folding the claim as true.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Perform a full versioned release — bump version, draft release notes, commit, tag, push, and create the GitHub release page. Reads `.claude/project.yaml` for the project's release-context and version-bump helpers; the skill text is universal, the helpers are project-supplied. Use when the user says "cut a release", "ship vX.Y.Z", "release", or after a shippable phase.
One repeatable pass that keeps fak's growing namespace CRYSTAL-CLEAR - each similar-sounding concept (cache / vCache / KV cache / cachemeta / the provider prompt-cache; guard vs gate; the two witnesses) given one canonical name, a written definition, and an explicit line drawn against the siblings it is confused with. Runs the concept-disambiguation scorecard (tools/concept_disambiguation_scorecard.py) over a data dir of confusable-concept rows cross-checked against the tree (the grounding token must appear in the production corpus; the glossary anchor must exist; a distinct_from reference must resolve), turns each HARD defect into a required fix (rename a true canonical collision, write a missing definition, draw + anchor a missing distinction, replace a fabricated grounding) and each uncovered tree token into a coverage row to ADD, retires disambiguation-debt worst-first, re-measures to PROVE the debt dropped, and commits only the scorecard lane by explicit path. The NAMING-clarity counterpart of conflation
One repeatable pass over README.md — the front door — that keeps ONLY the most important points current and honestly framed. Runs the freshness auditor (tools/readme_freshness_audit.py), turns each FAIL into a required edit and each WARN into a judgment call, applies the three front-page laws (SOTA-vs-us-never-naive, 6th-grade/Feynman-but-accurate, wide-audience), re-stamps the readme-verified marker, and commits ONLY README.md (+ the tool) by explicit path. Use after a release / version bump, after a headline number changes, when a link dies, or on a /loop cadence to keep the front page from rotting. The README's checking layer, the way curate-cluster is the index's.
Commit finished work cleanly on the shared trunk — lint the subject with `fak commit --preview`, then stage-and-commit EXACTLY your paths in one locked step via `fak commit --path … -m "…"`, verify the landed path-set and message are yours, and push when asked. Mechanizes the repo's "commit clean by default" mantra (trunk-only, explicit pathspec, DCO sign-off, Conventional-Commits subject with a bindable `(fak <leaf>)` stamp). Use when the user says "commit this", "ship my work", "commit cleanly", "land my change", or when green work is finished and ready to land on the shared trunk.
Flag content-redundant SKILL.md pairs as merge candidates — pairwise body cosine similarity over the skill pack, sorted highest-first. Use when the pack feels bloated or two skills seem to overlap, to find "skills A and B are N% token-similar — consider merging" candidates before a skill-lifecycle archive pass. Proposal only — never merges, edits, or deletes a skill.
The generic scoring doctrine the whole fak scorecard family instantiates — how to BUILD a new deterministic, tree-cross-checked, debt-driving scorecard and how to RUN any of them as a repeatable RSI pass. Every sibling (code-quality, docs, doc-appeal, seo, demo-quality, repo-hygiene, observability, learning, industry, agent-readiness, product, persona) is the same machine pointed at a different surface: pure KPIs over a data-dir or the git-tracked tree, cross-checked against reality so the score can't be gamed by editing data, folded into one *-debt integer + an A–F grade, emitting a control-pane payload, retired worst-first by ADDING the real thing, and proven with a before/after compare. Use when adding a NEW scorecard for a surface that isn't measured yet, when refining the scoring method itself, or when you need the shared contract behind quality-score / appeal-score / industry-score / agent-readiness / product-score / persona-score.
| name | verify |
| description | Bind a done-claim to a GREEN test run of the changed package, not just diff shape. Use after a commit claims a package/feature is done and you want to run that commit's affected tests and report CLAIM_TEST_GREEN / CLAIM_TEST_RED / CLAIM_TEST_UNRUN before folding the claim as true. |
dos commit-audit (and the dispatch-tick witness it feeds) grades whether a
commit's DIFF did the KIND of thing its subject claimed — diff-witnessed vs a
forgeable subject-only. That is commit SHAPE, explicitly not correctness: a
diff-witnessed commit can still fail its own tests. This skill runs the missing
rung — the resolving commit's affected-package tests — and binds the pass to
"done". It is additive to the shape witness, never a replacement.
The only path to a GREEN done-binding is a test that actually RAN and PASSED.
A runner that never fired, found no test-bearing changed package, or faulted is
CLAIM_TEST_UNRUN — a valid, surfaced state, never a pass. Do not fold a
done-claim as true on shape alone, and never let UNRUN masquerade as GREEN.
Take the commit whose done-claim you are checking (the resolving commit for the
issue, usually HEAD after a ship). Its changed .go files determine which
packages carry the test binding:
git show --name-only --pretty=format: <sha>
Map each changed file to its package dir; a changed dir with no *_test.go
carries nothing to bind a pass to — it contributes to UNRUN, not a vacuous pass.
Run the changed packages' tests through the repo's own affected-test seam, which
is WSL-aware — on native Windows it drives the tests through ./test.ps1, else
go test:
fak affected # runs the affected packages' tests through the WSL-aware seam
If you are checking one known package directly, the explicit equivalent is:
go test ./<changed-pkg>/... -count=1
Do not substitute a build (go build) or a vet for a test run — a green build
is shape, not the test binding.
Fold the run into exactly one claim, the same grader the dispatch-tick witness
records into the .witness sidecar's test_claim field:
| Ran? | Passed? | Claim | Meaning |
|---|---|---|---|
| yes | yes | CLAIM_TEST_GREEN | affected tests ran and passed — done-binding holds |
| yes | no | CLAIM_TEST_RED | affected tests ran and FAILED — a diff-witnessed commit can still be RED; this is the gap the rung closes |
| no | — | CLAIM_TEST_UNRUN | no resolving commit, no test-bearing changed package, or the runner was off/faulted — valid, surfaced, NOT a pass |
ran=false always outranks a stale passed bit: a run that never happened is
UNRUN, never GREEN.
On a live dispatch tick with the in-tick runner enabled, the shell records this
same binding for you. The default runner is opt-in via FAK_WITNESS_TEST_RUN
(unset -> the rung records UNRUN so the hot loop never eats an in-tick
go test); this verify skill is the always-on manual consumer of the same
rung. When a sidecar exists, read it rather than re-running:
# the per-worker resolve log's sibling: resolve-<N>-<stamp>.witness
# its `test_claim` key carries CLAIM_TEST_GREEN / _RED / _UNRUN
Preserve the rung alongside the shape verdict — a diff-witnessed +
CLAIM_TEST_GREEN is a strictly stronger done-binding than diff-witnessed
alone, and diff-witnessed + CLAIM_TEST_RED is a REFUTATION.
CLAIM_TEST_GREEN: the done-claim is test-bound; safe to fold as done.CLAIM_TEST_RED: do not fold as done; the change fails its own tests —
re-open or fix, and say so.CLAIM_TEST_UNRUN: do not fold as GREEN; report the missing test run and why
(no changed test-bearing package, runner off/faulted). UNRUN is honest, not a
pass.diff-witnessed as if it proved the tests pass.CLAIM_TEST_UNRUN read as "probably green".*_test.go as a pass (that is a vacuous run —
it is UNRUN).