메뉴
Full audit of the current package: comments, README and AGENTS.md. Runs check-comments, check-readme, check-agents-md and reflect in sequence.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Create, update or improve the local AGENTS.md for the current package. Use when setting up a new package, after significant code changes, or when new process knowledge should be captured. Imports `AGENTS-shared.md` and extends it with package-specific context.
Audit `AGENTS-shared.md` and all convention files for quality and correctness. Checks for internal consistency, gaps, redundancies and alignment between conventions and agent instructions.
Suite-level audit of skill files and guidelines. Runs snap-check-skills, snap-check-guidelines and reflect in sequence.
Audit all skill files for quality and consistency. Checks each skill against project conventions and checks across all skills for gaps, redundancies, conflicts, ambiguities, scope issues, stale references and naming discrepancies.
Review the current conversation for learnings that should improve existing skills, AGENTS.md files or memory. Use at the end of a session to capture non-obvious findings before they're lost.
Scan the whole project for improvements in implementation, architecture, API, clarity, refactoring and guideline compliance. Records findings as TODO comments in source or as bullets in the README TODO section.
| name | check |
| description | Full audit of the current package: comments, README and AGENTS.md. Runs check-comments, check-readme, check-agents-md and reflect in sequence. |
Run the following skills in order:
/check-comments — audit all source comments for typos, correctness, clarity, brevity and policy compliance/check-readme — verify the README against the public API and update it/check-agents-md — review and update the local AGENTS.md/reflect — capture any process improvements from the sessionComplete each skill fully before starting the next.
After all sub-skills complete, before running reflect, cross-check the audited content for mutual consistency:
Present any findings using the standard format defined in AGENTS.md.
/reflect because it runs as the final step.