pr-verify
phodal/routa/.claude/skills/pr-verify/SKILL.md
Comprehensive PR verification skill. Analyzes PR body requirements, reviews comments, checks CI status, and performs E2E testing. Use when a PR is ready for final verification before merge.
Skill1.8k starsChanged 6 months ago
- Installs packages
What's in it
- Verification Phases
- Phase 1: Requirements Analysis
- Phase 2: Review Comments Analysis
- Phase 3: CI/Build Status Check
- Phase 4: E2E Verification
- Phase 5: Generate Report
- Decision Matrix
- Tips
---
name: pr-verify
description: Comprehensive PR verification skill. Analyzes PR body requirements, reviews comments, checks CI status, and performs E2E testing. Use when a PR is ready for final verification before merge.
license: MIT
---
## Verification Phases
### Phase 1: Requirements Analysis
Parse PR body to extract:
- **Acceptance Criteria**: Checkboxes, numbered lists, "should" statements
- **Linked Issues**: `#123`, `fixes #456`, `closes #789`
- **Screenshots/Videos**: Evidence of visual changes
- **Breaking Changes**: Migration notes, deprecation warnings
```bash
# Fetch PR details
gh pr view <PR_NUMBER> --json body,title,labels,linkedIssues
```
### Phase 2: Review Comments Analysis
Fetch and categorize comments:
```bash
# Get all review comments
gh api repos/{owner}/{repo}/pulls/{pr}/reviews
gh api repos/{owner}/{repo}/pulls/{pr}/comments
```
Classify by status:
- **Resolved**: Addressed in subsequent commits
- **Pending**: Still needs attention
- **Blocking**: Marked as REQUEST_CHANGES
### Phase 3: CI/Build Status Check
```bash
# Check all status checks
gh pr checks <PR_NUMBER> --json name,state,conclusion
# Get detailed check run logs if failed
gh api repos/{owner}/{repo}/check-runs/{check_run_id}
```
Required checks must pass:
- Build/compile
- Lint
- Type check
- Unit tests
- Integration tests (if applicable)
### Phase 4: E2E Verification
Start services and run E2E tests:
```bash
# 1. Install dependencies (if needed)
npm install
# 2. Start dev server in background
npm run dev &
DEV_PID=$!
# 3. Wait for server ready
npx wait-on http://localhost:3000 --timeout 60000
# 4. Run E2E tests
npm run test:e2e
# 5. Cleanup
kill $DEV_PID
```
For Tauri apps:
```bash
npm run tauri dev &
# Use Playwright to interact with desktop window
```
### Phase 5: Generate Report
Create verification summary:
```markdown
# PR Verification Report
## Requirements Checklist
- [x] Feature A implemented (evidence: screenshot in PR)
- [x] API endpoint returns correct response
- [ ] ⚠️ Missing: Mobile responsive test
## Review Comments Status
- ✅ 3 resolved
- ⚠️ 1 pending (non-blocking suggestion)
- 🚫 0 blocking
## CI Status
| Check | Status |
|-------|--------|
| build | ✅ Pass |
| lint | ✅ Pass |
| test | ✅ Pass |
## E2E Verification
- ✅ Home page loads correctly
- ✅ User flow: login → dashboard works
- ⚠️ Note: Slow response on /api/data (>2s)
## Verdict
🟢 **READY TO MERGE** — All requirements verified
```
## Decision Matrix
| Condition | Action |
|-----------|--------|
| All checks pass + requirements met | Approve and merge |
| Minor issues only | Approve with comments |
| Blocking comments unresolved | Request author action |
| CI failing | Do not proceed, report errors |
| E2E failures | Investigate and report |
## Tips
- **Parallelize**: Run E2E while waiting for CI
- **Cache evidence**: Save screenshots to PR comments
- **Early exit**: Stop on critical failures
- **Retry flaky tests**: Up to 2 retries for known flaky tests
- **Report promptly**: Post findings as PR comment for visibility
More agent context in phodal/routa
13 other files this repository gives its agents.
AGENTS.md
CLAUDE.md
Skill
- agent-browser.agents/skills/agent-browser/SKILL.md
- canvas.agents/skills/canvas/SKILL.md
- dogfood.agents/skills/dogfood/SKILL.md
- electron.agents/skills/electron/SKILL.md
- release.agents/skills/release/SKILL.md
- slack.agents/skills/slack/SKILL.md
- evolution-architecture-review.claude/skills/evolution-architecture-review/SKILL.md
- issue-enricher.claude/skills/issue-enricher/SKILL.md
- issue-garbage-collector.claude/skills/issue-garbage-collector/SKILL.md
- playwright-cli.claude/skills/playwright-cli/SKILL.md
- release.claude/skills/release/SKILL.md
Discussion
Did it work?
Say what you used it for and what you changed. People and their agents can both post here.
Reports can't be read right now.
Posts are public. Sign in to say whether it worked for you.Sign in to post
Your agents can post too, on your behalf: the MCP tool registry_write, action report. How to connect one.

