AI 代码审查 — Claude Code 自动化 PR 审查 Skill
安装代码审查 Skill，将 Claude Code 变成对抗式代码审查员。自动化 PR 审查、安全检查、性能审计和风格规范——在人类审查员打开 PR 之前就发现 Bug。
安装 Code Reviewer Skill
# Install the Code Reviewer agent skill
curl -s https://api.tokrepo.com/raw/code-reviewer-agent \
> ~/.claude/skills/code-reviewer/SKILL.md
# Or install via TokRepo CLI
npx tokrepo install code-reviewer-agent
Bug Hunter — Adversarial AI Code Review + Auto-Fix
Bug Hunter is an adversarial code review skill that runs Hunter/Skeptic/Referee agents, reports confirmed issues, and supports canary-style auto-fixes.
Gemini CLI Extension: Code Review — Quality Checks
Gemini CLI extension for automated code review. Analyzes code quality, suggests improvements, and checks coding standards.
AI Code Review Checklist — Ship Better with AI Help
Structured checklist for reviewing AI-generated code before merging. Covers correctness, security, performance, maintainability, and AI-specific pitfalls like hallucinated imports and phantom APIs.
PR-Agent — AI-Powered Code Review for Pull Requests
AI code reviewer for GitHub/GitLab/Bitbucket PRs. Auto-generates descriptions, reviews code, suggests improvements, answers questions. By Qodo. 10.7K+ stars.
Continue — AI Code Review Agents for CI/CD
Continue runs AI agents as code review checks on PRs. 32.2K+ GitHub stars. Markdown-defined checks, pass/fail with diffs, VS Code + JetBrains. Apache 2.0.
Sourcery — Instant AI Code Reviews
Sourcery is an automated AI code reviewer for GitHub PRs plus an IDE assistant, helping teams reduce review time with summaries and inline suggestions.
Roo Code — AI Coding Agent with Custom Modes
Fork of Cline with custom agent modes, boomerang orchestration, and multi-model routing. Create specialized AI agents for coding, review, and architecture tasks.
Optio — Workflow Orchestrator for AI Coding Agents
Automates the full AI development lifecycle from task planning to merged PR. Orchestrates AI agents through planning, execution, code review, and merge. 800+ GitHub stars.
Codex Plugin for Claude Code — OpenAI Cross-Review
Official OpenAI plugin that adds Codex code review and task delegation inside Claude Code. Get adversarial reviews, rescue debugging, and multi-model collaboration.
Code Review Checklist — 系统化代码审查 Skill
涵盖安全、性能、可维护性、正确性和测试的系统化代码审查模式。包含严重等级分类、三遍审查流程、结构化反馈指南及需避免的反模式。来源 ClawHub/wpank
Claude Code Agent: Code Reviewer — Automated PR Review
Claude Code agent for comprehensive code reviews. Checks security, performance, maintainability, and best practices. Install with one command.
GitHub MCP — Repository Management for AI Agents
Official GitHub MCP server that gives AI agents access to repos, issues, PRs, code search, and Actions. Create branches, review PRs, and manage projects via natural language. 5,000+ stars.
Gitea — Lightweight Self-Hosted Git Service
Gitea is a painless self-hosted Git service with code hosting, review, CI/CD, package registry, and project management — a lightweight GitHub/GitLab alternative.
code-review-graph — MCP Context for Smarter Reviews
code-review-graph builds a Tree-sitter code graph and exposes minimal review context via MCP; verified 16,364★ and claims ~8.2× token reduction on 6 repos.
Trail of Bits Skills — Security Plugin Marketplace
Add Trail of Bits’ marketplace to run security code reviews, static analysis, and supply-chain checks via repeatable Claude Code plugins.
Claude Code Agent: Terraform Iac Reviewer
Terraform-focused agent that reviews and creates safer IaC changes with emphasis on state safety, least privilege, module patterns, drift detection, and plan/apply discipline
Claude Code Agent: Flutter GO Reviewer
|
Claude Code Security Review — PR Audit Action
Claude Code Security Reviewer is a GitHub Action that scans PR diffs for security issues and comments findings on the PR using a Claude API key.
Claude Code Agent: Se System Architecture Reviewer
System architecture review specialist with Well-Architected frameworks, design validation, and scalability analysis for AI and distributed systems
Claude Code Agent: AI Ethics Advisor
AI ethics and responsible AI development specialist. Use when reviewing an AI system for bias, fairness violations, or regulatory compliance gaps; when generating a model card,...
Compound Engineering — Skills & Agents for Better Reviews
Compound Engineering ships Claude/Codex skills for planning, review, and codifying knowledge; install the npm plugin and use /ce-* commands (16,691★).
Claude Code Agent: Code Reviewer
Use this agent when you need to conduct comprehensive code reviews focusing on code quality, security vulnerabilities, and best practices. Specifically:\\n\\n<example>\\nContext: Developer has submitted a pull request with changes to critical authentication...
Claude Code Agent: Hlbpa
Your perfect AI chat mode for high-level architectural documentation and review. Perfect for targeted updates after a story or researching that legacy system when nobody...
Claude Code Agent: Architect Reviewer
Use this agent when you need to evaluate system design decisions, architectural patterns, and technology choices at the macro level. Specifically:\ \ \ Context: Team has...
gstack — Claude Code Skill Pack (23 roles)
Install gstack in Claude Code to unlock 23 role-based slash commands for planning, review, QA, and shipping. Clone + run ./setup; team mode supported.
AI 代码审查：捕获人类遗漏的问题
The AI code reviewer is the single most impactful skill you can add to Claude Code. Human code reviewers are great at architectural feedback and design patterns — but they consistently miss edge cases, security vulnerabilities, performance regressions, and inconsistencies with existing code. An AI reviewer reads every line, checks every branch, and never gets tired or distracted.
The best code review skills work adversarially: they actively try to break your code rather than rubber-stamping it. They ask "what happens if this input is null?" and "what if this API call times out?" and "is this SQL query vulnerable to injection?" before the code reaches production. Install one of the review skills below, run claude review on your diff, and get a structured report of issues ranked by severity.
For teams, AI code review is the force multiplier that lets senior engineers focus on architecture while the AI handles line-by-line quality. Pair with AI testing tools for automated test generation on flagged code paths, or add security audit tools for deeper vulnerability scanning. For CI/CD integration, check the DevOps tools category — several tools run AI review as a GitHub Action on every PR.
A human reviewer spends 10 minutes on a 500-line PR. An AI reviewer spends 10 seconds — and catches the null pointer the human skimmed past.
常见问题
What is an AI code review skill?+
A SKILL.md file that teaches Claude Code how to perform structured code reviews. When installed, it gives the agent a specific methodology: check for security issues, identify performance problems, flag inconsistencies with the codebase, verify error handling, and produce a severity-ranked report. Unlike generic 'review this code' prompts, a skill encodes expert review patterns that apply consistently to every PR.
How do I run AI code review with Claude Code?+
After installing a code review skill, run Claude Code in your repo and ask: 'Review the changes in this PR' or 'Review the diff between main and this branch.' The skill instructs the agent to check security, performance, correctness, style, and testing — then produce a structured report. You can also configure Claude Code Hooks to run reviews automatically on every commit.
Can AI code review replace human reviewers?+
No — but it can handle 80% of the mechanical work. AI excels at: catching null pointer bugs, identifying missing error handling, flagging security vulnerabilities, checking naming conventions, and verifying test coverage. Humans are still better at: evaluating architecture decisions, assessing code readability, understanding business context, and mentoring junior developers. The best setup uses both.
Does AI code review work in CI/CD pipelines?+
Yes. Tools like Continue and CodeRabbit run AI-powered reviews as GitHub Actions on every pull request. They post review comments directly on the PR — no manual trigger needed. For Claude Code specifically, you can set up a CI job that runs 'claude review' on the PR diff and posts results as a comment. Browse TokRepo's DevOps tools for pre-built CI integrations.
What languages does AI code review support?+
Claude Code's code review skills work with any language — TypeScript, Python, Go, Rust, Java, C++, Ruby, and more. The AI understands language-specific patterns: it knows Go error handling conventions, Python type hints, TypeScript strict mode requirements, and Rust ownership rules. Review quality is highest for popular languages with large training data.