Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问clear审计通过

code-review代码审查

Agent Skill

code-review 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。

总安装

222

周安装

9

GitHub Stars

公开资料未说明

下载量

70
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

3

许可证

MIT

最后核验

2026-05-01

来源状态

来源可访问

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(代码审查)
来源仓库:https://github.com/vapvarun/claude-backup
仓库路径:skills/code-review
安装命令:
npx skills add https://github.com/vapvarun/claude-backup --skill code-review
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/vapvarun/claude-backup --skill code-review

简介

code-review 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。

  • 它支持根据关键词、任务场景或来源线索进行信息匹配,适用于研究类工作流。
  • 通过 npx skills add 命令从指定 GitHub 仓库安装并使用。
  • 安装前建议确认权限范围、维护状态及是否触发联网或文件操作。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Code Review

Systematic code review methodology for identifying issues, security vulnerabilities, and optimization opportunities.

Team Review Guidelines

Communication Style

When providing review feedback:

  1. Be friendly and constructive

- Good: "What do you think about using X here? It might help with Y." - Good: "Nice solution! One small suggestion..." - Avoid: "This is wrong." - Avoid: "You should have done X."

  1. Ask before assuming

- Good: "I'm curious about this approach - could you explain the reasoning?" - Avoid: "Why didn't you just...?"

  1. Acknowledge good work

- "Great catch on the edge case!" - "Clean and readable!" - "I learned something from this."

Double Verification Process

Every PR requires TWO passes:

Pass 1 - Security (Blocking)

  • All security checklist items MUST pass
  • Any security issue = Request Changes

Pass 2 - Quality (Advisory)

  • Standards, performance, maintainability
  • These are suggestions, not blockers (unless severe)

Review Phases

Phase 1: Preparation

  1. Understand Context

- Read PR description and linked issues - Understand the feature/fix being implemented - Check if there are related changes in other PRs - Review the commit history for context

  1. Identify Standards

- Check project's coding standards - Review existing patterns in the codebase - Note any specific requirements (performance, security)

Phase 2: Systematic Analysis

Work through each category methodically:

Security Review

Input Validation


// GOOD: Parameterized + validated app.post('/user', (req, res) => {const schema = z.object({id: z.number().positive()}); const {id} = schema.parse(req.body); db.query('SELECT * FROM users WHERE id =?', [id]);}); ```

### Authentication & Authorization

// GOOD: Verify ownership app.delete('/api/posts/:id', authenticate, async (req, res) => {const post = await Post.findById(req.params.id); if (post.authorId!== req.user.id &&!req.user.isAdmin) {return res.status(403).json({error: 'Forbidden'});} await post.delete();}); ```

Security Checklist

  • SQL/NoSQL injection prevented (parameterized queries)
  • XSS prevented (output encoding, CSP headers)
  • CSRF protection in place
  • Sensitive data not logged or exposed
  • API keys/secrets not hardcoded
  • File uploads validated and sanitized
  • Rate limiting on sensitive endpoints
  • Proper error messages (no stack traces in production)

Performance Review

N+1 Query Detection


// GOOD: Eager loading const posts = await Post.find().populate('author'); ```

### Performance Checklist

- No N+1 queries (use eager loading)
- Expensive operations cached appropriately
- Database queries use proper indexes
- Large datasets paginated
- Async operations properly awaited
- No unnecessary re-renders (React)
- Memory leaks prevented (cleanup handlers)
- Heavy computations memoized or offloaded

## Code Quality Checklist

- Clear, descriptive naming conventions
- Functions are single-purpose (< 30 lines ideal)
- Cyclomatic complexity reasonable (< 10)
- No code duplication (DRY principle)
- Proper error handling throughout
- Edge cases considered
- Magic numbers/strings extracted to constants
- Comments explain "why", not "what"

## Edge Cases Checklist

Always verify these scenarios are handled:

### Data Edge Cases

- Empty arrays/collections
- Null or undefined values
- Zero values (especially in division)
- Negative numbers where only positive expected
- Empty strings vs null vs undefined
- Unicode and special characters in strings

### User Input Edge Cases

- Missing required fields
- Fields with only whitespace
- Duplicate submissions (double-click)
- Concurrent updates to same resource
- Invalid date formats
- Timezone handling
- File upload with no file selected

### API Edge Cases

- Network timeout handling
- Rate limit exceeded responses
- API returns unexpected data structure
- Partial success in batch operations
- Graceful degradation when service unavailable

### WordPress-Specific Edge Cases

- Post with no featured image
- User with no display name
- Taxonomy with no posts
- Widget in empty sidebar
- Shortcode with no attributes
- Multisite subdirectory vs subdomain
- Translation missing for locale
- Cron job running during high traffic
- Option not yet saved (first install)
- Meta value of 0 vs meta not existing
- get_post() returning null
- WP_Query with no results
- Current user not logged in (ID = 0)

## Language-Specific Checks

### PHP/WordPress

- Escaping output (esc_html, esc_attr, etc.)
- Nonces for form submissions
- Capability checks for actions
- Prepared statements for queries
- WordPress coding standards followed
- Cron hooks: No name collision between cron hook and internal do_action()
- Cron scheduling: Events scheduled on activation, cleared on deactivation
- Hook callbacks: No recursive/self-triggering patterns
- Settings sanitization handles unchecked checkboxes
- Transient expiration set appropriately
- Object cache checked before expensive queries
- REST endpoints have permission_callback
- AJAX handlers verify nonce and capability

### JavaScript/TypeScript

- === used instead of ==
- Proper TypeScript types (no any abuse)
- async/await error handling
- No prototype pollution risks
- ESLint/Prettier rules followed

### React

- Hooks rules followed
- Keys provided for lists
- useEffect dependencies correct
- No state mutations
- Proper component composition

## Review Output Format

Structure feedback as:

### Critical Issues (Must Fix)

- Security vulnerabilities
- Data loss risks
- Breaking changes without migration
- Failing tests

### Warnings (Should Fix)

- Performance concerns
- Code maintainability
- Missing error handling

### Suggestions (Nice to Have)

- Refactoring opportunities
- Better naming
- Additional documentation

### Positive Notes

- Good patterns followed
- Clean code
- Thorough testing

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

04

需要参考平台分布和安装热度时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Claude Code

27.07%
按下载量换算19

windsurf

23.63%
按下载量换算17

Codex

16.57%
按下载量换算12

Gemini CLI

11.1%
按下载量换算8

OpenCode

7.69%
按下载量换算5

Antigravity

3.33%
按下载量换算2

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。

来源信息

继续浏览同类 Skills