메뉴
Turn a set of merged pull requests into a clean, grouped release-notes section. Triggers: user says 'write release notes', 'summarize this release', 'changelog for these PRs'.
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
| name | release-note-writer |
| description | Turn a set of merged pull requests into a clean, grouped release-notes section. Triggers: user says 'write release notes', 'summarize this release', 'changelog for these PRs'. |
| user-invocable | true |
| metadata | {"author":"Fixture Author","version":"1.0.0"} |
A small fixture skill used only as test material for skill-eval-runner evals. It is intentionally simple.
A markdown section titled ## <version> with one ### <group> subsection per non-empty group.
Write a clear Conventional Commits message from a set of staged changes. Triggers: user says 'write a commit message', 'draft a commit', or pastes a diff and asks how to describe it. Use whenever changes are staged and need a well-formed commit subject and body.
Generate evaluation test cases for any skill by browsing its SKILL.md and references. Produces evals.json with [auto]-prefixed expectations. Triggers: user says 'generate evals', 'create evals for skill', 'scaffold evals'; or after creating/updating a skill that needs test coverage. Use this skill whenever a skill exists but has no evals, or when evals need updating after skill changes.
Run skill evaluations by spawning sub-agents for test execution and LLM-as-judge grading. Includes blind A/B comparison and eval-viewer for visual review. Triggers: user says 'run evals', 'evaluate skill', 'test skill'; or after generating/updating a skill with evals.
Parallel codebase assessment with specialized subagents for QA, Technical Debt, Functional, and Architecture analysis. Use when analyzing a whole codebase for quality, performing code audits, assessing code health, identifying improvements, evaluating technical debt across a project, or when onboarding to a new codebase. Even if the user just says "review this project" or "what should we improve", this skill applies.
Parallel PR review with specialized subagents for QA, Technical Debt, Functional, and Architecture analysis. Use when reviewing pull requests, checking test quality, detecting tech debt, validating functional completeness, or verifying architectural consistency.
Triage a repository's open issue backlog by proposing priorities, labels, and comments. Detects already-triaged issues, identifies potential duplicates, and generates a gh-based apply script.
