Token导航 LogoToken导航

Agent Skill · 研究检索

code-reviewer(代码审查员)

code-reviewer 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。

code-reviewer 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,权限为接触敏感数据;许可证信息还不明确,第三方扫描有未通过项,按安装量在研究检索类的 73,675 个 Skill 里排第 72,323。

总安装 188周安装 8GitHub Stars 7下载 66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-reviewer(代码审查员)
来源仓库:https://github.com/9tykeshav/mern-ninja-cc
仓库路径:skills/code-reviewer
安装命令:
npx skills add https://github.com/9tykeshav/mern-ninja-cc --skill code-reviewer
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/9tykeshav/mern-ninja-cc --skill code-reviewer

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。

code-reviewer 用于代码审查,结合通用智能与 MERN 技术栈专长,优先评估项目健康状态。适合在 Codex、Claude、Cursor、Gemini CLI 中执行全面代码审查,识别构建问题、文档缺失和测试覆盖不足。通过检查编译错误、依赖项异常和文件规模,快速定位部署阻塞点,避免深入阅读低质量代码。

安装前请确认权限范围和维护状态,注意是否触发联网、命令执行或文件读写操作。建议结合来源仓库和 README 核验具体用法,确保审查流程符合项目实际需求。

来源仓库内路径:skills/code-reviewer

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/9tykeshav/mern-ninja-cc --skill code-reviewer;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

    它需要接触敏感数据,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-reviewer 怎么安装?

在终端执行 npx skills add https://github.com/9tykeshav/mern-ninja-cc --skill code-reviewer。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

code-reviewer 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-reviewer 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

code-reviewer 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-reviewer 用的人多吗?

公开数据是累计安装 188 次,近一周安装 8 次,下载 66 次,GitHub 7 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-reviewer 安全吗?

第三方扫描有未通过项,风险较高。除非你确认过源码,否则不建议直接安装到常用环境。

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。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 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。

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
code-reviewer

代码审查员 · 本页

1887接触敏感数据未明确未通过

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 Reviewer

Overview

Comprehensive code review: General intelligence + MERN specialization.

Philosophy: Check project health FIRST, then dive into code. A 6,000-line file is a problem regardless of what's in it.

Review Workflow

Phase 0: Project Health (Do This First)

Before reading any code, assess project health:

  1. Build status: Run tsc --noEmit or check for compilation errors
  2. Project docs: Read README, any STATUS/BUGS/TODO files - look for deployment blockers
  3. Test health: Do tests exist? Check package.json scripts, look for test directories
  4. File sizes: find src -name "*.ts" -o -name "*.tsx" | xargs wc -l | sort -n | tail -20
  5. Dependencies: Check for npm audit issues, unusual deps (Angular in React?)

Stop here if: Build is broken, docs say "DO NOT DEPLOY", or critical blockers found. Report immediately.

Phase 1: Scope Detection

  1. Identify scope from context:

- Full repo → Broad review, sample key files - Feature/PR → All changed files - Single file → Deep dive

  1. Detect layers: React? Express? MongoDB? Node.js?
  2. If ambiguous → ask user

Phase 2: Review by Priority

PriorityFocusSeverity
0. BlockersBuild failures, "DO NOT DEPLOY", broken deploysSTOP
1. SecurityInjection, auth, secrets, XSSCritical
2. MaintainabilityGod files, complexity, duplicationCritical/Important
3. PerformanceN+1, missing indexes, re-rendersImportant
4. TestingNo tests, low coverage, flaky testsImportant
5. Best PracticesError handling, async patternsSuggestion
6. ArchitectureAPI design, state managementSuggestion

Load reference files ON-DEMAND when you hit MERN-specific edge cases.

Phase 3: Report

Use the output format below. Offer to fix starting with Critical.

Output Format

# MERN Code Review

## Project Health
- Build: [Compiles / X errors / Not checked]
- Tests: [X passing / X failing / None found]
- Blockers: [Any deployment blockers from docs]
- Large files: [Files >500 lines]

## Scope
[What was reviewed]

## Summary
- Files reviewed: X
- Issues: X Critical, X Important, X Suggestions

## Critical (Must Fix)
### [C1] Category: Title
**File:** `path:line`
**Why:** [1-2 sentences]
**Fix:** [Code or instruction]

## Important (Should Fix)
### [I1] Category: Title
...

## Suggestions
- `file:line` - Note

## What's Good
- [Positive observations]

## Verdict
[Ready to deploy / Blocked / Needs fixes] - [1 sentence reason]

---
**Ready to fix these?** Starting with Critical issues.

Checklists

Minimum required checks. Report other issues you find during review.

Blockers (Check First)

  • Project compiles without errors
  • No "DO NOT DEPLOY" or similar warnings in docs
  • No critical security advisories in npm audit

Security

  • No $where, $ne, $regex with user input (NoSQL injection/ReDoS)
  • No dangerouslySetInnerHTML without DOMPurify
  • JWT in httpOnly cookies, not localStorage
  • Secrets in env vars, not hardcoded (check config files too, not just code)
  • Helmet middleware configured
  • CORS properly restricted
  • Rate limiting on auth endpoints
  • Input validation on all endpoints
  • No eval() or new Function() with user input

Maintainability

  • No file >500 lines (god files)
  • No function >50 lines
  • No class/component with >20 methods
  • No deep nesting (>4 levels)
  • No copy-paste blocks >10 lines (DRY)
  • Clear naming (no cryptic abbreviations)
  • Consistent code style

Performance

  • No N+1 queries (use populate/$lookup)
  • Indexes on frequently queried fields
  • .lean() for read-only Mongoose queries
  • No fs.readFileSync in request handlers
  • React.memo on expensive components
  • useCallback/useMemo where beneficial
  • Pagination on list endpoints

Testing

  • Tests exist for critical paths (auth, payments, core flows)
  • Test coverage reasonable (>50% for services)
  • No skipped/commented-out tests
  • Tests actually assert behavior (not just "doesn't crash")
  • Mocks don't hide real integration issues

Best Practices

  • Async errors handled (try/catch or error middleware)
  • useEffect cleanup functions present
  • No floating promises (unhandled async)
  • Middleware order correct (body-parser before routes, error handler last)
  • Environment variables validated at startup
  • Graceful shutdown handlers

Architecture

  • Consistent API response format
  • Service layer between controllers and DB
  • Types aligned frontend/backend
  • No circular dependencies
  • Clear module boundaries
  • No god components (React >300 lines)
  • State management appropriate for complexity

Red Flags (Immediate Critical)

These are automatic Critical issues:

  • eval(), new Function() with user input
  • Hardcoded secrets/credentials in code
  • dangerouslySetInnerHTML without sanitization
  • JWT/auth tokens in localStorage
  • Missing auth middleware on protected routes
  • $where clause with user input
  • File >1000 lines
  • "DO NOT DEPLOY" in project docs
  • npm audit critical vulnerabilities

Scope Calibration

ScopePhase 0Code DepthFocus
Single fileSkipDeepAll checklists on that file
Last commitQuickMediumChanged lines + immediate context
Feature/PRQuickMediumAll changed files
Full repoFullBroadSample key files, architecture

Reference Files

Load ONLY when you encounter MERN-specific patterns you need to verify:

When to LoadReference
NoSQL query security questionsecurity.md
React hooks/re-render issuereact.md
Express middleware questionexpress.md
MongoDB schema/index questionmongodb.md
Node.js async/memory issuenodejs.md
API design/auth flow questionfullstack.md

Do NOT load all references upfront. They're for edge cases, not general review.

Don't

  • Don't claim "no issues found" without actually searching for them
  • Don't report on code you haven't read
  • Don't classify style issues as Critical

Examples

God File Detection

Found: EventService.ts - 6,165 lines
→ Critical [C1] Maintainability: God file
→ Recommend split into: EventQueryService, EventBookingService,
   EventGuestService, EventInviteService (~500 lines each)

Missing Health Check

Found: CURRENT_STATUS_AND_BUGS.md contains "DO NOT DEPLOY"
→ Critical [C1] Blocker: Deployment blocked by known issues
→ Fix TypeScript errors in EditEventModal.tsx before proceeding

Security + Specific Fix

Found: No Helmet middleware in index.ts
→ Critical [C2] Security: Missing security headers
→ Fix: npm install helmet && app.use(helmet())

平台分布

Codex

38.14%
按下载量换算25

Claude

29.45%
按下载量换算19

Cursor

19.24%
按下载量换算13

Gemini CLI

10.17%
按下载量换算7

安全审计

Gen Agent Trust Hub

未通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。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