code-review-skill(1) — User Commands (en

  ___ ___  ___  ___        ___ _____   _____ _____      __    ___ _  _____ _    _    
 / __/ _ \|   \| __|  ___ | _ \ __\ \ / /_ _| __\ \ /\ / / __/ __| |/ /_ _| |  | |   
| (_| (_) | |) | _|  |___||   / _| \ V / | || _| \ V  V /__\__ \ ' < | || |__| |__ 
 \___\___/|___/|___|      |_|_\___| \_/ |___|___| \_/\_/    |___/_|\_\___|____|____|

$ man code-review-skill LANG=zh_CN | en_US

v1.0 · awesome-skills · MIT · 20 languages · 16,000+ lines

NAME

code-review-skill — A comprehensive, modular code review skill for Claude Code

SYNOPSIS

Use code-review-skill to review this PR
Use code-review-skill to review this <component>
Use code-review-skill for [security | performance | architecture] review

DESCRIPTION

A production-grade code review skill. It transforms AI-assisted code review from vague suggestions into a structured, consistent, expert-level collaborative process.

Core is only ~190 lines; the full 16,000+ lines of language guides load on demand. Covers 20+ mainstream languages and frameworks — progressive loading, zero overhead.

Every finding carries an explicit severity label. Every review proceeds through four phases: PR context · high-level assessment · line-by-line analysis · summary & decision.

LANGUAGES

┌── frontend ──┘

react.mdReact 19, Hooks, Server Components, TanStack v5 .................870

vue.mdVue 3.5, Composition API, Composables, Watchers .................920

angular.mdAngular 17+, Signals, Standalone, Zoneless ..........................420

svelte.mdSvelte 5, Runes, SvelteKit, SSR/CSR boundaries ..................1,060

typescript.mdTypeScript strict mode, generics, immutability ..................540

css-less-sass.mdCSS/Less/Sass variables, responsive, compatibility ..............660

┌── backend ──┘

python.mdPython async, typing, pytest, mutable defaults .................1,070

django.mdDjango/DRF security, N+1, serializers, async views ..............1,030

java.mdJava 17/21, Spring Boot 3, virtual threads, JPA ................800

php.mdPHP 8.x, types, PDO, security, Composer ...........................700

go.mdGoroutines, channels, context, interface design .................990

rust.mdOwnership, async/await, unsafe, cancellation safety .............840

csharp.mdC# 12 / .NET 8, EF Core, ASP.NET Core, LINQ .....................520

nestjs.mdNestJS DI, guards, interceptors, DTO validation .................590

┌── mobile / systems ──┘

kotlin.mdKotlin/Android coroutines, Compose, Flow, null safety ...........1,020

swift.mdSwift 5.9+/6, SwiftUI, concurrency, Sendable, optionals ..........930

c.mdC pointer safety, undefined behavior, resources .................210

cpp.mdC++ RAII, Rule of 0/3/5, move semantics, noexcept ...............300

qt.mdQt object model, signals/slots, GUI performance .................190

┌── cross-cutting ──┘

architecture-review-guide.mdSOLID, anti-patterns, coupling ...470

performance-review-guide.mdWeb Vitals, N+1, complexity .......850

code-quality-universal.mdTOCTOU, leaky abstractions, sprawl ....320

security-review-guide.mdInjection, XSS, secrets, all langs .....—

PHASES

  ┌─────────────┐     ┌─────────────┐     ┌─────────────┐     ┌─────────────┐
  │   context   │ ─▶  │ high level  │ ─▶  │ line by line│ ─▶  │   decide    │
  │    2-3m     │     │    5-10m    │     │   10-20m    │     │    2-3m     │
  └─────────────┘     └─────────────┘     └─────────────┘     └─────────────┘

1. context gathering — 2-3 min: Read the PR description and linked issues, assess scope, check CI status, understand the business intent.
2. high-level review — 5-10 min: Evaluate architectural fit, performance impact, file organization, test strategy. See the whole first.
3. line-by-line analysis — 10-20 min: Logic correctness · security · performance · maintainability · edge cases. One by one.
4. summary & decision — 2-3 min: Summarize findings, name what was done well, deliver approve / comment / request-changes.

SEVERITY

● [blocking] must fix — resolve before merge; security / correctness / serious logic
● [important] should fix — strongly recommended; discuss if you disagree
● [nit] nice to have — style or preference; non-blocking
● [suggestion] alternative — worth considering; author decides
● [learning] educational — no action needed; share knowledge
● [praise] good work — say it out loud when you see it

INSTALLATION

Clone into the Claude Code skills directory. Two commands.

# macOS / Linux
$ git clone https://github.com/awesome-skills/code-review-skill.git \
    ~/.claude/skills/code-review-skill

# Windows PowerShell
PS> git clone https://github.com/awesome-skills/code-review-skill.git `
       "$env:USERPROFILE\.claude\skills\code-review-skill"

EXAMPLES

Use code-review-skill to review this PR runs the full four-phase review
Review this React component loads react.md · checks Hooks · Server Components
Security review of this Go service loads go.md + security-review-guide.md together
Architecture review loads the architecture guide · SOLID · anti-patterns · coupling

FILES

~/.claude/skills/code-review-skill/
├── SKILL.md                              # core, loaded on activation (~190 lines)
├── README.md
├── LICENSE                               # MIT
├── reference/                            # on-demand language guides
│   ├── react.md vue.md angular.md ...
│   └── architecture-review-guide.md ...
├── assets/
│   ├── review-checklist.md              # quick reference
│   └── pr-review-template.md            # PR comment template
└── scripts/
    └── pr-analyzer.py                   # PR complexity analyzer

AUTHORS

awesome-skills: maintainer, primary author
contributors: see graphs/contributors

COPYRIGHT

Copyright (c) 2025 awesome-skills.
Released under the MIT License.
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.

CODE-REVIEW-SKILL(1) awesome-skills CODE-REVIEW-SKILL(1)