菜单
Classify failing tests as flaky vs. real regressions and propose a next step for each. Invoke when the user pastes failing test output, asks to triage CI failures, or asks "are these tests flaky".
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
Scaffold a new HTTP handler with service function, store method (if needed), and table-driven test, following the project's internal/{api,service,store} layout. Invoke when the user asks to add an endpoint, handler, or route.
Scaffold a new API endpoint with request/response Pydantic schemas, service function, route handler, and a pytest test — following the project's layered layout. Invoke when the user asks to add an endpoint, route, or API method.
Scaffold a new React component with matching test file and consistent folder layout, following the project's conventions in src/features or src/components. Invoke when the user asks to create, scaffold, or add a new component.
Write React Testing Library tests for an existing component, following the project's RTL conventions (prefer role/label queries, mock at network boundary, no act()). Invoke when the user asks to test, add tests for, or cover a component.
Run this repo's documentation and shell quality gates (shellcheck, markdownlint, lint-claude-md) the same way CI runs them. Invoke before opening a PR, or when the user asks to lint, validate, or pre-flight the docs.
Generate or update CHANGELOG.md with an entry for the unreleased changes since the last git tag. Invoke when the user asks to cut a release, update the changelog, or summarize changes since v-something.
| name | test-triage |
| description | Classify failing tests as flaky vs. real regressions and propose a next step for each. Invoke when the user pastes failing test output, asks to triage CI failures, or asks "are these tests flaky". |
| allowed-tools | Bash(git log:*), Bash(git blame:*), Read, Grep, Glob |
Given a batch of failing tests, classify each and propose one concrete action per test.
Prefer in this order:
git log -5 --follow <test-file> and the same for the implementation. Look for changes in the last ~24h or the last PR.A table, one row per failing test:
| Test | Classification | Likely cause | Proposed action |
|---|
Then, below the table, list the tests ranked by fix-first priority (real regressions > stale tests > environmental > flaky-with-high-frequency > low-frequency flakes).
unknown and say what you'd need.