Token导航 LogoToken导航

Agent Skill · 研究检索

pr-quality-checklist(公关质量检查表)

pr-quality-checklist 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

pr-quality-checklist 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为操作浏览器;许可证为 MIT,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 25,971。

总安装 3,660周安装 151GitHub Stars 40下载 1,196核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:pr-quality-checklist(公关质量检查表)
来源仓库:https://github.com/bobmatnyc/claude-mpm-skills
仓库路径:skills/pr-quality-checklist
安装命令:
npx skills add https://github.com/bobmatnyc/claude-mpm-skills --skill pr-quality-checklist
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/bobmatnyc/claude-mpm-skills --skill pr-quality-checklist

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

提供高质量拉取请求的综合指南与最佳实践清单。pr-quality-checklist 属于研究检索类 Skill,可作为该场景下的辅助能力补充。包含作者与审稿人模板、大小控制建议与标准化规范。适用于团队 PR 文化建设、新人培训与审核效率提升。通过 github 方式安装,需指定正确仓库与技能路径。

checklist 内容为通用建议,具体项目需结合实际情况调整。

来源仓库内路径:skills/pr-quality-checklist

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/bobmatnyc/claude-mpm-skills --skill pr-quality…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要操作浏览器,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

pr-quality-checklist 怎么安装?

在终端执行 npx skills add https://github.com/bobmatnyc/claude-mpm-skills --skill pr-quality-checklist。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

pr-quality-checklist 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

pr-quality-checklist 需要什么权限?

权限级别是操作浏览器。该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

pr-quality-checklist 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

pr-quality-checklist 用的人多吗?

公开数据是累计安装 3,660 次,近一周安装 151 次,下载 1,196 次,GitHub 40 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

pr-quality-checklist 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
pr-quality-checklist

公关质量检查表 · 本页

3,66040操作浏览器MIT通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

PR Quality Checklist Skill

Summary

Comprehensive guide for creating high-quality pull requests that are easy to review, well-documented, and follow team standards. Includes templates, size guidelines, and best practices for both authors and reviewers.

When to Use

  • Before creating any pull request
  • When reviewing others' PRs
  • To establish team PR standards
  • During onboarding of new team members
  • When PR reviews are taking too long

Quick Checklist

Before Creating PR

  • PR title follows convention: type(scope): description
  • Description includes summary, related tickets, and test plan
  • Changes are focused and related (single concern)
  • Code is self-reviewed for obvious issues
  • Tests added/updated and passing
  • No debugging code (console.log, debugger, etc.)
  • TypeScript/type errors resolved
  • Documentation updated if needed
  • Screenshots included for UI changes
  • Changeset added (for user-facing changes)

PR Title Format

Convention

{type}({scope}): {short description}

Types

  • feat: New feature
  • fix: Bug fix
  • refactor: Code change that neither fixes a bug nor adds a feature
  • docs: Documentation only
  • style: Formatting, missing semicolons, etc. (no code change)
  • test: Adding or updating tests
  • chore: Maintenance tasks, dependency updates
  • perf: Performance improvement
  • ci: CI/CD changes

Examples

feat(auth): add OAuth2 login support
fix(cart): resolve race condition in checkout
refactor(search): extract query param parsing logic
docs(api): update authentication examples
test(payment): add integration tests for Stripe
chore(deps): update Next.js to v15
perf(images): implement lazy loading for gallery
ci(deploy): add staging environment workflow

Scope Guidelines

  • Use component/module name: auth, cart, search, api
  • Be specific but not too granular: user-profile not user-profile-settings-modal
  • Consistent with codebase structure

PR Description Template

Standard Template

## Summary
<!-- 1-3 bullet points describing what changed -->
- Added OAuth2 authentication flow
- Integrated with Auth0 provider
- Updated login UI components

## Related Tickets
<!-- Link to Linear/Jira/GitHub issues -->
- Closes #123
- Related to #456
- Fixes ENG-789

## Changes
<!-- Detailed list of changes -->
### Added
- `src/lib/auth/oauth2.ts` - OAuth2 client implementation
- `src/app/api/auth/callback/route.ts` - Auth callback handler
- `src/components/OAuthButtons.tsx` - OAuth login buttons

### Modified
- `src/components/LoginForm.tsx` - Added OAuth buttons
- `src/lib/auth/index.ts` - Export new auth methods
- `README.md` - Updated setup instructions

### Removed
- `src/lib/auth/legacy.ts` - Deprecated auth code
- `src/components/OldLoginForm.tsx` - Replaced by new form

## Screenshots
<!-- Required for UI changes -->
### Desktop
![Desktop view](https://imgur.com/desktop.png)

### Mobile
![Mobile view](https://imgur.com/mobile.png)

### Before/After (if applicable)
| Before | After |
|--------|-------|
| ![Before](before.png) | ![After](after.png) |

## Testing
<!-- How was this tested? -->
- [ ] Unit tests added/updated
- [ ] Integration tests pass
- [ ] Manual testing completed
- [ ] Tested on Chrome, Firefox, Safari
- [ ] Mobile responsive (tested on iOS/Android)
- [ ] Edge cases tested (empty state, error states, loading)

## Breaking Changes
<!-- If any breaking changes -->
⚠️ **Breaking Change**: The `login()` function signature has changed.

**Before:**

login(username: string, password: string)


**After:**

login({ username: string, password: string, provider?: 'local' | 'oauth' })


**Migration:**

// Old await login('user', 'pass');

// New await login({ username: 'user', password: 'pass' });


## Checklist

- Code follows project style guidelines
- Self-review completed
- Comments added for complex logic
- Documentation updated (if needed)
- Changeset added (if user-facing change)
- No console.log/debugger statements left
- No TypeScript errors
- Tests pass locally
- Build succeeds

Minimal Template (for small changes)

## Summary
Brief description of the change.

## Changes
- Modified `file.ts` to fix XYZ

## Testing
- [ ] Tests pass
- [ ] Verified manually

Size Guidelines

Ideal PR Size

  • Lines changed: < 300 (excluding generated files)
  • Files changed: < 15
  • Review time: < 30 minutes
  • Commits: 1-5 logical commits

Size Categories

SizeLinesFilesReview TimeRecommendation
XS< 501-35 min✅ Ideal for hotfixes
S50-1503-815 min✅ Good size
M150-3008-1530 min⚠️ Consider splitting
L300-50015-251 hour❌ Should split
XL500+25+2+ hours❌ Must split

When to Split PRs

1. By Feature Phase
PR #1: MVP implementation (core functionality)
PR #2: Polish and edge cases
PR #3: Additional features
2. By Layer
PR #1: Database schema changes
PR #2: Backend API implementation
PR #3: Frontend UI integration
PR #4: Tests and documentation
3. By Concern
PR #1: Refactoring (no behavior change)
PR #2: New feature (builds on refactored code)
PR #3: Tests for new feature
4. By Risk Level
PR #1: High-risk changes (need careful review)
PR #2: Low-risk changes (routine updates)

Exceptions to Size Limits

  • Generated code (migrations, API clients, types)
  • Renaming/moving files (show with git mv)
  • Bulk formatting changes (separate PR, pre-approved)
  • Third-party library integration (well-documented)

Refactoring PRs

Refactoring Template

## Summary
Refactoring and cleanup for [area]

**Goal**: Improve code maintainability without changing behavior

## Motivation
- Reduce code duplication (3 similar functions → 1 reusable utility)
- Improve type safety (any → specific types)
- Remove dead code identified during feature work

## Related Tickets
- ENG-XXX: Improve [area] maintainability
- ENG-YYY: Remove deprecated [feature]

## Stats
- **Lines added**: +91
- **Lines removed**: -1,330
- **Net**: -1,239 ✅
- **Files changed**: 23

## Changes

### Removed (Dead Code)
- `/api/old-endpoint` - unused, replaced by `/api/new-endpoint` in v2.0
- `useDeprecatedHook.ts` - replaced by `useNewHook.ts` (ENG-234)
- `legacy-utils.ts` - functions no longer called anywhere

### Refactored
- **Extracted common logic**: Query param parsing now in `lib/url-utils.ts`
- **Consolidated validation**: 3 duplicate Zod schemas → 1 shared schema
- **Improved types**: Replaced 12 `any` types with proper interfaces

### No Behavior Changes
- [ ] All tests pass (no test modifications needed)
- [ ] Same inputs → same outputs
- [ ] No user-facing changes

## Testing
- [ ] Full test suite passes (no new tests needed)
- [ ] Manual smoke test completed
- [ ] No regressions identified

## Review Focus
- Verify removed code is truly unused (checked with `rg` search)
- Confirm refactored logic is equivalent
- Check no subtle behavior changes introduced

Refactoring Best Practices

  • Separate refactoring from features: Never mix
  • Verify zero behavior change: Tests should not need updates
  • Document removed code: Explain why safe to remove
  • Search thoroughly: Use rg/grep to verify no usage
  • Celebrate negative LOC: Highlight code reduction

Review Checklist

For Authors (Self-Review)

Before requesting review, check:

Code Quality
  • Code is DRY (no obvious duplication)
  • Functions are single-purpose and focused
  • Variable names are clear and descriptive
  • No magic numbers or hardcoded values
  • Error handling is comprehensive
  • Edge cases are handled
Testing
  • Tests cover new functionality
  • Tests are meaningful (not just for coverage)
  • Edge cases are tested
  • Error conditions are tested
  • No flaky tests introduced
Documentation
  • Complex logic has comments
  • Public APIs have JSDoc/docstrings
  • README updated (if setup changed)
  • Migration guide (if breaking changes)
Performance
  • No obvious performance issues
  • Database queries are efficient (indexes considered)
  • No N+1 queries
  • Large datasets handled appropriately
Security
  • No hardcoded secrets
  • Input validation present
  • Authorization checks in place
  • No SQL injection vulnerabilities
  • Sensitive data not logged

For Reviewers

First Pass (5 minutes)
  • PR description is clear
  • Changes align with stated goal
  • Size is reasonable (< 300 LOC preferred)
  • No obvious red flags
Code Review (15-30 minutes)
  • Logic is correct
  • Edge cases are handled
  • Error handling is appropriate
  • Code is readable and maintainable
  • No performance issues
  • Security considerations addressed
Testing Review
  • Tests cover the changes
  • Tests are meaningful
  • No flaky tests
  • Edge cases tested
Documentation Review
  • PR description matches changes
  • Comments explain "why" not "what"
  • Breaking changes documented
  • Migration guide provided (if needed)
Approval Criteria
✅ Approve: Code is good, minor nits only
💬 Comment: Suggestions but not blockers
🔄 Request Changes: Issues must be fixed before merge

Common Mistakes

1. Vague PR Titles

❌ "Fix bug"
❌ "Updates"
❌ "WIP"
❌ "Changes from code review"

✅ "fix(auth): prevent duplicate login requests"
✅ "feat(cart): add coupon code support"

2. Missing Context

❌ "Changed the function"
Why? What was wrong? What's the impact?

✅ "Refactored parseQueryParams to handle nested objects

Previously failed when query params contained dots (e.g., user.name).
Now correctly parses nested structures using qs library.

Fixes ENG-456"

3. Too Large

❌ 2,000 lines changed across 50 files
❌ "Implement entire authentication system"

✅ Split into:
  - PR #1: Add database schema for auth (150 LOC)
  - PR #2: Implement JWT utilities (100 LOC)
  - PR #3: Create login endpoint (200 LOC)
  - PR #4: Add login UI (250 LOC)

4. Mixed Concerns

❌ Single PR with:
  - Feature implementation
  - Dependency updates
  - Refactoring
  - Bug fixes
  - Formatting changes

✅ Separate PRs:
  - PR #1: feat(feature)
  - PR #2: chore(deps)
  - PR #3: refactor(component)

5. No Screenshots for UI Changes

❌ "Updated the header design"
(No screenshots)

✅ "Updated the header design"
[Desktop screenshot]
[Mobile screenshot]
[Before/After comparison]

6. Incomplete Testing Notes

❌ "Tested locally"
How? What scenarios? What browsers?

✅ "Testing completed:
- Unit tests: All 47 tests pass
- Manual testing: Tested login flow on Chrome, Firefox, Safari
- Edge cases: Tested expired tokens, invalid credentials, network errors
- Mobile: Verified on iOS Safari and Android Chrome"

7. Leaving Debug Code

❌ console.log('user data:', user);
❌ debugger;
❌ // TODO: fix this later
❌ import { test } from './test-utils'; (unused)

✅ Clean code without debugging artifacts

8. No Changeset (for User-Facing Changes)

❌ New feature shipped without changelog entry

✅ .changeset/new-feature.md:
---
"@myapp/web": minor
---

Add OAuth2 login support with Auth0 integration

Summary

Quick Reference

PR Title Formula
{type}({scope}): {clear, concise description}
Ideal PR Characteristics
  • Size: < 300 LOC, < 15 files
  • Focus: Single concern, related changes only
  • Documentation: Clear description, test plan, screenshots (if UI)
  • Quality: Self-reviewed, tested, no debugging code
  • Timeline: Ready to merge, not WIP
Review Mindset

As Author:

  • Make reviewer's job easy
  • Provide context and reasoning
  • Test thoroughly before requesting review
  • Respond promptly to feedback

As Reviewer:

  • Be constructive and specific
  • Focus on correctness and maintainability
  • Ask questions when unclear
  • Acknowledge good practices
Red Flags
  • 🚩 No description or "WIP" title
  • 🚩 Over 500 LOC changed
  • 🚩 Mixed concerns (feature + refactor + deps)
  • 🚩 No tests or only changing tests
  • 🚩 "Trust me, it works"
  • 🚩 Console.log/debugger statements
  • 🚩 TypeScript errors ignored
  • 🚩 No screenshots for UI changes

Use this skill to maintain high PR quality standards and streamline code review processes.

平台分布

Claude Code

30.32%
按下载量换算363

Antigravity

21.61%
按下载量换算258

Gemini CLI

15.82%
按下载量换算189

OpenCode

11.74%
按下载量换算140

Cursor

7.06%
按下载量换算84

windsurf

3.27%
按下载量换算39

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。

继续浏览同类 Skills