菜单
Analyze current branch changes vs main and determine whether e2e tests need to be created or updated. Produces a structured report with coverage gaps and recommendations. Use after completing a feature or before creating a PR.
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
Create a git worktree and prime the development environment for a new feature or refactor. Use when starting isolated feature work or needing a separate working copy.
Diagnose and fix a failed GitHub Actions job run. Given a run URL/ID (or the latest failed run on the current branch), pulls the failure via the gh CLI, reproduces it in an isolated worktree off the exact failing commit, fixes it, and verifies locally. Use when the user says "fix the failing CI", "the GitHub Action failed", invokes "/fix-ci", or points at a failed GitHub Actions run.
Reviews code for bugs, security vulnerabilities, and quality issues. MUST be invoked automatically after completing a coding task (feature, bug fix, or refactor) - do not wait for user to request it. Also use when the user asks to review changes, check a diff, or look over a branch/PR before committing.
Reflect on the current session to find mistakes, user corrections, ambiguous instructions, and failures — then propose durable root-cause fixes to the project's instruction docs (CLAUDE.md, .claude/rules, .claude/skills) so they don't recur. Use when the user asks to reflect on the session, run a retro/postmortem, capture lessons, or after a session with notable corrections or repeated mistakes.
Create Knex database migrations for the Benefriches API. Use when adding, modifying, or removing database columns/tables. Handles schema changes (create table, add/drop/rename columns), data migrations, and updates to tableTypes.d.ts.
Run Benefriches end-to-end tests with automatic Docker stack lifecycle management. Starts the e2e docker-compose stack, waits for readiness, runs Playwright tests, and tears down the stack on success. Use when the user asks to run e2e tests, end-to-end tests, or invokes /run-e2e-tests. Accepts an optional argument to focus on specific tests (e.g., tests/create-site/ or a specific spec file).
| name | e2e-coverage-check |
| description | Analyze current branch changes vs main and determine whether e2e tests need to be created or updated. Produces a structured report with coverage gaps and recommendations. Use after completing a feature or before creating a PR. |
| effort | medium |
| allowed-tools | Agent |
| user-invocable | true |
Analyze whether current branch changes require new or updated e2e tests.
Launch the e2e-coverage-checker agent with this prompt:
Analyze the current branch's changes against main and produce an e2e coverage check report. Follow your instructions step by step: collect the diff, classify files, run the deep pass where needed, map coverage, and generate the report.
Relay the agent's report to the user exactly as produced. Do not summarize or modify it.
