Token导航 LogoToken导航

Agent Skill · 开发

code-review-practices(代码审查实践)

code-review-practices 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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。

code-review-practices 是开发类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在开发类的 21,746 个 Skill 里排第 16,391。

总安装 509周安装 21GitHub Stars 6下载 166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
unknown

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:code-review-practices(代码审查实践)
来源仓库:https://github.com/hack23/homepage
仓库路径:skills/code-review-practices
安装命令:
npx skills add https://github.com/hack23/homepage --skill code-review-practices
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。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/hack23/homepage --skill code-review-practices

这个 Skill 是做什么的

开发类 Skill 面向写代码、改代码和排查问题,一般会读写仓库文件,部分还会执行构建或测试命令。

Code Review Practices 确立尊重性沟通、及时响应和彻底检查三大核心原则指导审查行为。适用于注重工程文化和知识共享的团队,强调审查过程本身的教育价值。提供具体操作指南如 24 小时内响应、小 PR 快速处理等可量化执行的标准。需结合团队实际情况调整严格程度,避免教条化执行影响开发效率。

code-review-practices 属于开发类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/code-review-practices

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/hack23/homepage --skill code-review-practices;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

    它需要需要联网,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

code-review-practices 怎么安装?

在终端执行 npx skills add https://github.com/hack23/homepage --skill code-review-practices。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

code-review-practices 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。

code-review-practices 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

code-review-practices 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

code-review-practices 用的人多吗?

公开数据是累计安装 509 次,近一周安装 21 次,下载 166 次,GitHub 6 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。

code-review-practices 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。
  • 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

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 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。
  • 需要处理开发类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

开发类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
skill-vetter

技能审查员 · OpenClaw

8,779,018986操作浏览器MIT-0通过
ontology

本体论 · OpenClaw

6,636,960576只读MIT-0通过
entra-app-registration

内部应用程序注册 · Codex、Claude

6,631,872764接触敏感数据未明确通过
entra-app-registration

内部应用程序注册 · Codex、Claude

3,757,680201接触敏感数据未明确通过
nano-banana-pro

纳米香蕉专业版 · OpenClaw

3,526,898364接触敏感数据MIT-0有提醒
code-review-practices

代码审查实践 · 本页

5096需要联网未明确通过

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

Code Review Practices Skill

Purpose

Establishes effective code review practices that improve code quality, share knowledge, catch bugs early, and maintain security standards while fostering collaborative development culture.

Rules

Code Review Principles (MUST Follow)

Core Principles:

1. Respectful Communication
   - Focus on code, not the person
   - Ask questions, don't make demands
   - Provide constructive suggestions
   - Acknowledge good work

2. Timeliness
   - Review within 24 hours
   - Small PRs reviewed faster
   - Block releases for critical issues only
   - Don't let PRs go stale

3. Thoroughness
   - Understand the context
   - Test the changes locally
   - Check for side effects
   - Verify all review points

4. Continuous Learning
   - Explain reasoning behind feedback
   - Share knowledge and best practices
   - Learn from others' approaches
   - Iterate on review process

What to Review

MUST CHECK:

1. Functionality
   ✓ Code solves the stated problem
   ✓ Edge cases handled
   ✓ Error handling present
   ✓ No obvious bugs
   ✓ Follows requirements

2. Design & Architecture
   ✓ Appropriate abstraction level
   ✓ Follows project patterns
   ✓ No unnecessary complexity
   ✓ Proper separation of concerns
   ✓ Scalable approach

3. Code Quality
   ✓ Readable and maintainable
   ✓ Follows style guide
   ✓ DRY (Don't Repeat Yourself)
   ✓ Appropriate naming
   ✓ No magic numbers
   ✓ Commented where needed

4. Testing
   ✓ Unit tests included
   ✓ Integration tests if needed
   ✓ Test coverage adequate
   ✓ Tests are meaningful
   ✓ Edge cases tested

5. Security
   ✓ No hardcoded secrets
   ✓ Input validation present
   ✓ SQL injection prevention
   ✓ XSS protection
   ✓ Authentication/authorization checks
   ✓ Secure dependencies

6. Performance
   ✓ No obvious inefficiencies
   ✓ Database queries optimized
   ✓ Caching where appropriate
   ✓ No memory leaks
   ✓ Async operations used properly

7. Documentation
   ✓ README updated if needed
   ✓ API documentation updated
   ✓ Complex logic explained
   ✓ Breaking changes noted

Review Process

Pull Request Author:

Before Requesting Review:
1. Self-review your changes
2. Run tests locally
3. Run linter/formatter
4. Write clear PR description
5. Link related issues
6. Tag appropriate reviewers
7. Mark as draft if work-in-progress

PR Description Template:
## Changes
- [Bullet list of changes made]

## Why
[Explanation of motivation]

## Testing
- [ ] Unit tests added/updated
- [ ] Integration tests added/updated
- [ ] Manual testing completed

## Screenshots (if UI changes)
[Before/after screenshots]

## Checklist
- [ ] Code follows style guide
- [ ] Tests passing
- [ ] Documentation updated
- [ ] No secrets committed

Reviewer:

Review Steps:
1. Read PR description and linked issues
2. Understand the context and goals
3. Review code changes in logical order
4. Check out branch and test locally
5. Run automated checks (tests, linter)
6. Leave inline comments for specific issues
7. Provide summary comment
8. Approve, Request Changes, or Comment

Comment Types:
- **MUST FIX**: Blocking issue (security, bug, breaking)
- **SHOULD FIX**: Important but not blocking
- **NIT**: Nitpick, style preference
- **QUESTION**: Clarification needed
- **SUGGESTION**: Alternative approach
- **PRAISE**: Good work, clever solution

Feedback Guidelines

Effective Comments:

❌ BAD: "This is wrong."
✅ GOOD: "This could cause a race condition when multiple users
         access the cache simultaneously. Consider using a lock
         or atomic operation."

❌ BAD: "Why did you do it this way?"
✅ GOOD: "Have you considered using Array.map() here? It might be
         more readable and is functionally equivalent."

❌ BAD: "This is terrible naming."
✅ GOOD: "Could we rename 'tmp' to 'temporaryUserData' for clarity?
         It would help future maintainers understand the purpose."

Praise Effectively:

✅ "Great use of dependency injection here! Makes this very testable."
✅ "I like how you extracted this into a separate function. Much cleaner."
✅ "Excellent test coverage of edge cases!"
✅ "This optimization is impressive. Nice catch!"

Code Review Anti-Patterns

AVOID:

1. Rubber Stamping
   - Approving without thorough review
   - "LGTM" without understanding changes
   - Skipping testing locally

2. Nitpicking Without Value
   - Focusing only on style (use automated tools)
   - Personal preferences without reasoning
   - Arguing about subjective choices

3. Review Hostility
   - Hostile or condescending tone
   - Dismissing author's approach without discussion
   - Making it personal

4. Perfectionism
   - Holding up PR for minor issues
   - Requesting unnecessary refactors
   - Scope creep in review

5. Review Overload
   - PRs too large to review effectively
   - Too many reviewers required
   - Review fatigue

Pull Request Size Guidelines

MUST:

Target Sizes:
- Small: < 200 lines changed (ideal)
- Medium: 200-500 lines (acceptable)
- Large: 500-1000 lines (split if possible)
- XL: > 1000 lines (must split)

For Large Changes:
1. Break into smaller, logical PRs
2. Use feature flags for incremental merges
3. Create base branch for related PRs
4. Document overall design in RFC/issue

Exceptions (acceptable large PRs):
- Generated code
- Configuration updates
- Dependency bumps
- Data migrations

Security-Focused Review

MUST CHECK:

Authentication & Authorization:
✓ Authentication required for protected endpoints
✓ Authorization checks before resource access
✓ No role/permission bypass vulnerabilities
✓ Session management secure

Input Validation:
✓ All user input validated
✓ SQL injection prevention (parameterized queries)
✓ XSS prevention (output encoding)
✓ Path traversal prevention
✓ File upload validation

Sensitive Data:
✓ No secrets in code
✓ No API keys committed
✓ Sensitive data encrypted at rest
✓ PII handled per GDPR
✓ Secure logging (no passwords/tokens in logs)

Dependencies:
✓ npm audit / pip-audit passed
✓ No known vulnerable dependencies
✓ Dependencies from trusted sources
✓ Lock files updated

Example Security Comment:
"⚠️ SECURITY: This endpoint allows unauthenticated access to user
data. We need to add authentication middleware and verify the user
has permission to access this resource. See access-control SKILL.md
for implementation pattern."

Automated Review Checks

CI/CD Integration:

# .github/workflows/pr-checks.yml
name: PR Checks

on: [pull_request]

jobs:
  quality:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      - name: Lint
        run: npm run lint

      - name: Unit Tests
        run: npm run test:unit -- --coverage

      - name: Security Scan
        run: npm audit --audit-level=moderate

      - name: Code Coverage
        uses: codecov/codecov-action@v4
        with:
          fail_ci_if_error: true
          min_coverage: 80

      - name: CodeQL Analysis
        uses: github/codeql-action/analyze@v3

      - name: Lighthouse CI
        run: npm run lighthouse:ci

Branch Protection Rules:

MUST ENFORCE:
- ✓ Require pull request reviews (at least 1)
- ✓ Require status checks to pass
- ✓ Require branches to be up to date
- ✓ Require conversation resolution
- ✓ No force push
- ✓ No deletion
- ✓ Include administrators

Review Approval Process

Approval Criteria:

✅ APPROVE when:
- All MUST FIX issues resolved
- Tests passing
- Security checks pass
- No outstanding questions
- Code meets quality standards

💬 COMMENT when:
- Need clarification
- Have suggestions but not blocking
- Want to share knowledge
- Acknowledge good work

🔄 REQUEST CHANGES when:
- Critical bugs found
- Security vulnerabilities
- Breaking changes without discussion
- Tests missing or failing
- Major design concerns

Review Metrics & Improvement

TRACK:

Review Efficiency:
- Time to first review
- Time to merge after approval
- Number of review rounds
- PR size distribution

Review Quality:
- Bugs caught in review
- Bugs found in production
- Test coverage trends
- Security issues prevented

Team Health:
- Review participation distribution
- Review comment sentiment
- Knowledge sharing occurrences
- Team satisfaction with process

Examples

Exemplary Review Comments

Functionality Issue:

🐛 **Bug**: This will fail when `user.profile` is null. We should add
a null check or use optional chaining:

const email = user.profile?.email || 'no-email@example.com';

This handles the null case gracefully and prevents runtime errors.

Performance Suggestion:

⚡ **Performance**: This loops through users twice (filter then map).
We could combine into a single pass:

const activeUserEmails = users
  .filter(u => u.active)
  .map(u => u.email);

Could become:

const activeUserEmails = users.reduce((emails, u) => {
  if (u.active) emails.push(u.email);
  return emails;
}, []);

Though honestly, for small arrays the original is more readable. What's
the expected size of the users array?

Security Issue:

🔒 **SECURITY**: This SQL query is vulnerable to injection:

const query = `SELECT * FROM users WHERE email = '${email}'`;

We must use parameterized queries:

const query = 'SELECT * FROM users WHERE email = $1';
const result = await db.query(query, [email]);

See our secure-development SKILL.md for more examples.

Praise:

✨ Really nice abstraction here! Extracting this logic into a reusable
utility function makes it testable and DRY. The naming is also very
clear. Great work! 🎉

PR Description Template

## Summary
[Brief description of what this PR accomplishes]

## Related Issues
Closes #123
Related to #456

## Changes Made
- Added user authentication middleware
- Implemented JWT token validation
- Added unit tests for auth flow
- Updated API documentation

## Type of Change
- [ ] Bug fix
- [x] New feature
- [ ] Breaking change
- [ ] Documentation update

## Testing Performed
- [x] Unit tests (95% coverage)
- [x] Integration tests
- [x] Manual testing on staging
- [ ] E2E tests (not applicable)

## Screenshots (if applicable)
[Before/After screenshots for UI changes]

## Checklist
- [x] Code follows style guide
- [x] Self-reviewed code
- [x] Commented complex code
- [x] Updated documentation
- [x] No new warnings
- [x] Added tests
- [x] All tests passing
- [x] No secrets committed
- [x] CHANGELOG updated (if applicable)

## Deployment Notes
[Any special deployment steps or configuration changes needed]

## Rollback Plan
[How to rollback if issues arise in production]

Related Policies

Related Documentation

Tools

Code Review:

  • GitHub Pull Requests
  • GitLab Merge Requests
  • Gerrit
  • Phabricator

Automated Checks:

  • GitHub Actions
  • CodeQL
  • SonarQube
  • Codecov

Code Quality:

  • ESLint, Prettier
  • pylint, black
  • RuboCop
  • Checkstyle

平台分布

Codex

34.33%
按下载量换算57

Claude

33.1%
按下载量换算55

Cursor

18.82%
按下载量换算31

Gemini CLI

9.56%
按下载量换算16

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills