メニュー
Add a feature to this repository the safe way — spec, locate via maps, respect Stability, surgical implementation, verification, knowledge update. Use whenever the user asks to add, build, or implement functionality.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Bootstrap the ai/ knowledge layer for a repo that has none — draft the MODULE_MAP, diagrams, and guide docs as [inferred] for a human to audit, touching no source code. Use when the maps are still placeholders or the user asks to cold-start, bootstrap, or onboard a repository.
Bootstrap the ai/ knowledge layer for a repo that has none — draft the MODULE_MAP, diagrams, and guide docs as [inferred] for a human to audit, touching no source code. Use when the maps are still placeholders or the user asks to cold-start, bootstrap, or onboard a repository.
Deep, judgement-based adversarial code audit — hunts for stale cross-references, unescaped interpolation, platform gaps, generated-file ownership conflicts, and cross-module consistency rot that no mechanical check can catch. Read-only; writes a dated findings report.
Kit-maintainer diagnostic (LOCAL-ONLY, not distributed) that audits the README and docs/ for staleness against the repo's real command/skill/CLI roster, stamped tree, and cross-doc consistency. Read-only; produces a severity-ranked findings report.
Run verification and drift checks on the AI knowledge-base to identify missing documentation or stale references.
Deep-mine the source code to discover implemented features; writes ai/analysis/FEATURE_CATALOG.md.
| name | add-feature |
| description | Add a feature to this repository the safe way — spec, locate via maps, respect Stability, surgical implementation, verification, knowledge update. Use whenever the user asks to add, build, or implement functionality. |
The contract: no code before a spec, no edits to frozen code, no "done" without green tests, no merge without a review and a knowledge update.
If ai/lab/specs/SPEC_<name>.md doesn't exist, draft it from SPEC_TEMPLATE.md
(goal, scope, touch list, acceptance criteria, verification). Get the user's OK.
ai/guide/MODULE_MAP.md → which modules; note Stability of every target.ai/analysis/FEATURE_CATALOG.md → the "where new code lives" decision tree and
the 3-file rule for related features.ai/lab/WORKLOG.md → recent history: was this area just changed? is the
behavior deliberate?Any file in the touch list with Stability frozen or ? ⇒ stop and ask the human.
Record their approval in the spec before proceeding.
Make the smallest diff that satisfies the spec. Conventions per
ai/guide/CONVENTIONS.md — see also reference/checklist.md. Match the license
headers of neighboring files. No drive-by refactors, no layout changes, no
dependency additions unless the plan says so.
Run the narrowest suite that covers the change first, then the suite(s) the spec names. Report command run · exit status · failures verbatim. Red or unrun ⇒ not done.
FEATURE_MAP entry; catalog amendment; MODULE_MAP if layout changed; all [inferred].
Tell the user which tags await their [verified] flip.
Request /review-change on the diff (fresh context — not this session). Append
the work's row to ai/lab/WORKLOG.md linking spec, review, and commits, and move the
feature's ai/lab/ROADMAP.md row from Planned to Shipped (cross-link spec, WORKLOG
row, and PR by feature ID).