一键导入
ship
Create a PR with test evidence and performance verification
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。
菜单
Create a PR with test evidence and performance verification
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
Dual-perspective code review — Lead Dev (architecture/correctness) + Dev (implementation/tests). Runs entirely in the main agent for easy follow-up.
Update EggPdf documentation — README, docs/ wiki pages, site/, llms.txt, CLAUDE.md — to reflect current implementation
Implement a new feature following EggPdf's test-first, performance-aware development loop
Render HTML through EggPdf and analyze PDF output for visual correctness issues
Render all WebUI templates through the PDF pipeline and verify output correctness
Run EggPdf benchmarks and analyze performance for specific rendering scenarios
| name | ship |
| description | Create a PR with test evidence and performance verification |
| disable-model-invocation | true |
| user-invocable | true |
| allowed-tools | Bash, Read, Grep, Glob |
Create a pull request for the current feature branch, including test results and performance evidence.
$ARGUMENTS -- optional PR title override. If empty, auto-generate from branch name and commits.Pre-flight checks (ALL must pass):
dotnet build --configuration Release # No warnings
dotnet test --configuration Release # All green
If either fails, STOP. Fix first.
Verify we are NOT on main:
git branch --show-current
If on main, STOP and tell the user to create a feature branch.
Run performance check if any non-test code changed:
cd benchmarks/EggPdf.Benchmarks && dotnet run -c Release -- --filter *Render* --job short --exporters markdown
Analyze commits on this branch vs main:
git log main..HEAD --oneline
git diff main...HEAD --stat
Count test results:
dotnet test --configuration Release --verbosity minimal 2>&1 | tail -5
Create PR using gh pr create:
## Summary
- What was added/changed (1-3 bullets)
## Test Evidence
- X tests pass (Y new, Z existing)
- Test types: unit / layout / visual / E2E
## Performance
| Scenario | Time | Memory | Target | Status |
|----------|------|--------|--------|--------|
(include benchmark results if available)
No performance regressions detected.
## Checklist
- [ ] Tests written FIRST, then implementation
- [ ] All tests pass (new + existing)
- [ ] No performance regression
- [ ] Public API has XML documentation
- [ ] Multi-target compatible (netstandard2.0+)
Generated with [Claude Code](https://claude.com/claude-code)
Return the PR URL to the user.
main directly