← スキル一覧に戻る

qa-verification
by joelklabo
markdowntown scan CLI
⭐ 0🍴 0📅 2026年1月19日
SKILL.md
name: qa-verification description: Unified QA verification for both web UI and CLI work. Use when tasks require QA evidence, test execution, or verification checklists.
QA Verification
Overview
Use this skill to gather QA evidence. Choose the section that matches the work type (web UI vs CLI). If both apply, follow both sections.
Evidence Format
- Record evidence in a
## QA Evidencesection on the task. - Include: test command(s), screenshots/trace paths (if any), DevTools checks, logs reviewed, and CI status.
Web UI Verification
Required evidence:
- Playwright (or repo E2E) tests for affected UI flows.
- Screenshots of key UI states (store paths in task Files + QA evidence).
- Chrome DevTools evidence: console + network clean (no errors/warnings) and Issues panel checked.
- Frontend + backend logs reviewed with no errors/warnings.
- Playwright trace artifact captured on failure or flake (retain-on-failure or on-first-retry).
- CI command(s) run with green results.
Workflow:
- Identify affected UI flows and pages.
- Run Playwright (or repo-specified E2E) for those flows.
- Use resilient locators and web-first assertions if tests are added/updated (avoid manual visibility checks).
- Capture screenshots for key states; store in the repo’s standard screenshot path.
- Use DevTools (MCP if available) to check console, network, and Issues panel.
- Review FE/BE logs; fix any errors/warnings.
- If tests fail or are flaky, collect trace artifacts and record the path.
- Record QA evidence in the task before closing.
Fallback:
- If DevTools MCP is unavailable, capture Playwright trace/screenshots and browser console logs from the test run, and note the limitation.
Automated confirmation gate:
- Any UI change must complete the Web UI Verification steps before moving on.
- Do not wait for user approval; verification is automated evidence.
CLI Verification
Required evidence:
- Command transcripts: command + stdout + stderr + exit code.
- Non-interactive behavior: flags like
--no-input, stdin via pipe, and-for stdin/stdout when supported. - Output formats:
--jsonand plain output validated; schema stable if applicable. - Config precedence and location checks (flags/env/project/user/system; XDG if used).
- CI command(s) run with green results.
Workflow:
- Identify primary commands and flags touched by the change.
- Run happy-path and failure-path commands; capture exit codes and streams.
- Validate non-interactive paths (piped stdin,
-for stdin/stdout). - Verify output formats and schema stability.
- Verify config precedence and storage locations.
- Record QA evidence in the task before closing.
Guardrails
- If repo QA instructions conflict with this skill, follow repo rules.
- If the work spans both UI and CLI, apply both checklists.
スコア
総合スコア
60/100
リポジトリの品質指標に基づく評価
✓SKILL.md
SKILL.mdファイルが含まれている
+20
✓LICENSE
ライセンスが設定されている
+10
○説明文
100文字以上の説明がある
0/10
○人気
GitHub Stars 100以上
0/15
○最近の活動
3ヶ月以内に更新がある
0/10
○フォーク
10回以上フォークされている
0/5
✓Issue管理
オープンIssueが50未満
+5
✓言語
プログラミング言語が設定されている
+5
○タグ
1つ以上のタグが設定されている
0/5
レビュー
💬
レビュー機能は近日公開予定です