Claude code review skills and AI code review tools for Claude Code and Codex: PR review bots, local CLI reviewers, security review and multi-agent review. Every one is security-graded before you install it.
Of the 185 Claude Code Review Skills & Tools here, our pick after testing them is awesome-skills/code-review-skill (★ 1.9k). If you want a review skill at all: Found all eight planted defects with no false alarm on the clean pull requests, in about a minute and under two findings per pull request. That equals Claude Code with nothing installed: it adds a written checklist and a fixed report format, not detection. See the test →
On 2026-10-10 we ran 16 of these tools and 13 ran. Each reviewed the same 12 pull requests against a small Python service in a throwaway sandbox: 8 carry one planted defect (an off-by-one, an SQL injection, a card token in the logs, a missing permission check, a retry that never ends, a swallowed error, an inverted condition, a weakened test) and 4 are clean. Two numbers per tool, as in SWR-Bench: defects found, and blocking findings on the clean ones. A judge (gpt-6-astra, three passes) marked each review. Claude Code with no tool installed did the same job as a control.
Know what these do not look for: brooks-lint (design decay only; it passed a logged card token at 100/100); bug-hunter (ignores test files, so a weakened test goes through).
Whichever you pick, the hardest defect was not in the code: four of 13 reviewers passed a pull request that quietly weakened a test's assertion. Read test changes yourself.
Listed by defects found, then time. This set does not rank the tools at the top: Claude Code with no tool installed found all 8 planted defects in a minute with no false alarm, and six tools did the same. Tools that bring their own model reviewed with gpt-6-astra, the skills with Claude Opus 5.5, so a gap between those two groups may be the model.
What we found: No tool beat the control. Claude Code with nothing installed found all 8 defects in about a minute with no false alarm (1.8 findings per pull request); 6 tools matched it and none did better. What separates the tools is noise (0.5 to 7.2 findings per pull request) and time (1 to 30 minutes). The defect missed most was the weakened test: 4 of 13 let it through. See the control run →
| # | Tool | Defects found (of 8) | False alarms on 4 clean PRs | Findings per PR | Time | Reviewed by | Evidence |
|---|---|---|---|---|---|---|---|
| 1 | code-review-skill ★ 1,914 | 8/8 | 0 | 1.8 | 1 min | Claude Opus 5.5 | The reviews |
| 2 | guard-skills ★ 1,153 | 8/8 | 0 | 1.4 | 2 min | Claude Opus 5.5 | The reviews |
| 3 | review-forge ★ 220 | 8/8 | 0 | 1.3 | 2 min | Claude Opus 5.5 | The reviews |
| 4 | clean-architecture-skills ★ 90 | 8/8 | 0 | 2.7 | 2 min | Claude Opus 5.5 | The reviews |
| 5 | spencermarx/open-code-review ★ 331 Several reviewer personas write one report: every defect is in it, among 7 findings per pull request. |
8/8 | 0 | 7.2 | 11 min | Claude Opus 5.5 | The reviews |
| 6 | roborev ★ 1,746 Runs in the background on every commit; here it reviewed with gpt-6-astra through OpenCode and still found all eight. |
8/8 | 0 | 0.9 | 13 min | gpt-6-astra | The reviews |
| 7 | adamsreview ★ 243 Hit the 30-minute limit with one pull request unfinished; it found the defect in all seven it finished. |
7/7 | 0 | 0.5 | 30 min | Claude Opus 5.5 | The reviews |
| 8 | bug-hunter ★ 487 Treats test files as context only, by design, so it does not report a weakened test. |
7/8 | 0 | 0.8 | 3 min | Claude Opus 5.5 | The reviews |
| 9 | alibaba/open-code-review ★ 43,912 It skipped the two pull requests that only change tests, so the weakened test went unreviewed. |
7/8 | 0 | 0.6 | 7 min | gpt-6-astra | The reviews |
| 10 | vet ★ 519 | 7/8 | 0 | 1.1 | 18 min | gpt-6-astra | The reviews |
| 11 | brooks-lint ★ 1,506 Reviews for design decay (coupling, complexity), not security: the logged card token and the weakened test scored 100/100. |
5/8 | 0 | 0.9 | 3 min | Claude Opus 5.5 | The reviews |
| 12 | gentleman-guardian-angel ★ 1,180 It checks code against a rules file you write. The test repo had none, so the agent wrote one from gga's Python example; its 30 failures on clean pull requests are that file's docstring and type-hint rules, not false alarms. |
5/8 | see note | 6.0 | 9 min | gpt-6-astra | The reviews |
| 13 | warden ★ 414 Its verification pass threw out candidate findings on the pull requests it missed, the SQL injection among them. |
4/8 | 0 | 0.5 | 25 min | gpt-6-astra | The reviews |
An AI code reviewer is judged by two numbers that pull against each other: how many real problems it finds, and how much noise it adds. On SWR-Bench, 1,000 manually verified pull requests from GitHub, current review tools find too little and do best on functional bugs; running several reviews and combining them raised F1 by up to 43.67%. Noise has a price too. In a company deployment reported at ICSE 2025, 73.8% of the AI's comments were resolved, yet the average time to close a pull request rose from 5 hours 52 minutes to 8 hours 20 minutes, and engineers named faulty, unnecessary and irrelevant comments as the drawback. ByteDance's reviewer filters its own comments in a second stage to reach 75.0% precision, and measures adoption by whether the code a comment pointed at was then changed. So we judge a Claude Code review skill or tool by what it does about both: does it filter before it posts, and can you see whether its comments were acted on.
| Kind | How it works | Where it helps | What to watch |
|---|---|---|---|
| Code review skills 30 · codex-plugin-cc, code-review-skill |
A skill or plugin that has Claude Code or Codex review a diff on request. | Runs before you commit, inside the agent you already use. | The same model that wrote the code reviews it; ask a second model for important changes. |
| Local and CLI review 65 · open-code-review, code-review-graph |
A command-line reviewer you run on a branch or a diff, often with fixed rules plus an LLM. | Deterministic checks catch the known patterns; the model handles the rest. | Local runs leave no shared record of what was flagged and fixed. |
| PR review bots and CI 39 · pr-agent, shippie |
Comments on every pull request, in GitHub or GitLab, as a bot or a CI step. | Every change gets a first pass, and resolved comments give you an adoption rate. | Unfiltered comments slow merges; in one deployment, closing a PR took 2.5 hours longer on average. |
| Multi-agent review 26 · pr-af, chorus |
Several reviewers or models look at the same change and their findings are merged. | Combining reviews raised F1 by up to 43.67% on SWR-Bench. | Several times the tokens and the wait. |
| Security review 16 · DeepAudit, claude-code-security-review |
Agents that look for vulnerabilities, some of which reproduce and fix them in a sandbox. | Goes deeper on one class of problem than a general reviewer. | A complement to a general review, not a replacement; check the grade, since these tools run code. |
| Code review platforms 9 · px0, kodus-ai |
A self-hosted service that runs reviews across a team's repositories. | One place for rules, history and adoption numbers. | Someone has to own the rules and tune what gets posted. |
The research measures review comments on real pull requests and in company deployments, not the specific tools listed here. Until we run them ourselves, the questions below are how we read that research.
| Before you commit, on your own | A review skill in Claude Code or Codex, or a local CLI |
| Every pull request on a team | A PR bot that filters its comments and lets you set severity and scope |
| An important or risky change | Multi-agent review, or a second model on top of the first |
| Code that handles secrets, auth or user data | A security reviewer in addition to the general one |
The floor is install safety. 7 entries on this page are flagged CAUTION or UNSAFE, 3 only for how they install. Check before installing: fullcase-web (sudo_usage, sensitive_env_vars); codebase-argus (sensitive_env_vars); claude-audit (eval_usage); codesense-ai (tunnel_service).
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in multi-language ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.
🚀 PR Agent: The Original Open-Source PR Reviewer. This project is not the Qodo free tier.
DeepAudit: Open-source multi-agent code vulnerability auditing with one-click setup, sandbox PoC checks, Ollama deployment, reports, relay support.
An AI-powered security review GitHub Action using Claude to analyze code changes for security vulnerabilities.
A modular, stack-agnostic toolkit for AI coding agents to autonomously find, reproduce, and patch vulnerabilities.
A comprehensive code review skill for Claude Code, covering React 19, Vue 3, Rust, TypeScript, TanStack Query v5, and more.
Instant AI code reviews
px0 is an IDE built for reviewing AI-generated code, optimized for speed. It turns your browser into a zero-latency console with native Git and GitHub integrations, instant search across massive codebases, and seamless handoff to local AI coding harnesses.
Continuous background code review database for agents, work faster and smarter with accountability for every line of generated code.
AI code reviews grounded in 12 classic engineering books — decay risk diagnostics with book citations, severity labels, and 6 analysis modes including full-sweep auto-fix
Free, Micro AI Code Reviews That Run on Git Commit
😇 Gentleman Guardian Angel (gga) - Provider-agnostic code review using AI. Use Claude, Gemini, Codex, Ollama to enforce your coding standards.
Guard skills for coding agents, quality gates that catch AI-generated failure modes in code, tests, and docs
```bash
npx skills add amElnagdy/guard-skills --list
```AI Code Reviewer: Enhance your GitHub workflow with AI-powered code review! Get intelligent feedback and suggestions on pull requests using OpenAI's GPT-4 API, improving code quality and saving developers time.
Review your coding agent's diff in a terminal pane and send line comments back to Claude Code, Codex, OpenCode or Pi. A herdr plugin.
Mac app to review the code written by you or your agent with Claude in the loop.
A staged code-review workflow and local dashboard built with TypeSafe Jev.
🚀 AI-powered code review tool for GitHub, GitLab, Bitbucket Cloud, Bitbucket Server, Azure DevOps and Gitea — built with LLMs like OpenAI, Claude, Gemini, Ollama, Bedrock, OpenRouter and Azure OpenAI
Multi-LLM peer review for code decisions. Bring your own CLI; Chorus convenes 2-4 other LLMs to review the work before you ship.
Find issues worth your attention.
Local AI PR workbench — Claude/Codex review in isolated git worktrees, human-gated GitHub comments
Adversarial AI bug hunter with auto-fix skill for Claude Code, Cursor, Codex CLI, GitHub Copilot CLI, Kiro CLI, Opencode, Pi Coding Agent, and more. Multi-agent pipeline finds security vulnerabilities, logic errors, and runtime bugs — then fixes them autonomously on a safe branch.
Open-source Agentic code review tool inspired by DevinReview, using Recursive Language Models (RLM)
Open source agentic SAST. The engine behind hundreds of disclosed zero-days. 100+ AI security agents, any repo or PR diff, bring your own model.
An AI-powered GitHub code review tool that uses LLMs to detect high-confidence, high-impact issues—such as security vulnerabilities, bugs, and maintainability concerns.
Your code is under new management. Agents that review your code.
```bash
# Initialize warden in your repository
npx @sentry/warden init
# Add the built-in baseline reviews
npx @sentry/warden add security-review
npx @sentry/warden add code-review
# Run a pre-review on current branch changes
# Uses Pi. Set WARDEN_OPENAI_API_KEY, or WARDEN_ANTHROPIC_API_KEY for AnAgent skill for checking PR review comments, status checks, and description completeness
GitHub Action that reviews pull request diffs with an LLM and posts the comments on the PR
```yaml
# .github/workflows/code-review.yml
on:
pull_request:
types: [opened, synchronize]
permissions:
contents: read
pull-requests: write
jobs:
review:
runs-on: ubuntu-latest
steps:
- uses: cirolini/genai-code-review@v3
with:
api_key: ${{ secrets.GEMINSelf-hosted AI code reviewer with indexed PR reviews, walkthroughs, vulnerability scanning, dependency graphs, custom rules, and a learning loop.
AI-powered multi-agent code review. Simulates a customizable team of Engineers performing code review with built-in discourse.
🔮 Codeball – AI Code Review that finds bugs and fast-tracks your code
Get codebase-aware feedback in your GitHub, GitLab, Bitbucket pull requests or directly in your IDE.
Multi-lens code audit tool — 280 expert AI agents for code review, security testing, and infrastructure auditing
Security scanning skill for Claude Code, Codex and Antigravity. Finds the 21 most common vulnerabilities in AI-written code and shows how to fix each one.
A viewer for reviewing local code changes in small individual chapters. Works with any AI agent.
Bonk is a code & docs agent that can build, review and triage for you, built on Cloudflare Workers & OpenCode.
A GitHub Action that automatically reviews pull requests using Google's Gemini AI.
Multi-lens code review pipeline for Claude Code: deep review (Claude or Codex), auto-fix loop, interactive walkthrough, external-finding injection.
A Code Review tool for developers with comparisons, comment suggestions, change summaries, and code quality alerts to understand commits and find issues faster.
Local-first MCP plugin for continuous software-quality review by AI coding agents, powered by Jev.
Multi-LLM cross-review workflow for Claude Code. /work orchestrates plan/implement/bug-fix with 3 different model families voting on every gate.
review-forge is an Agent Skill for structured, auditable code review workflows
AI-powered cybersecurity code review skill for Claude Code. 8 specialist agents, OWASP 2025, CWE Top 25, MITRE ATT&CK, 11 languages, zero configuration.
Entity-level code review for Git. Graph-based risk scoring, change classification, commit untangling. 95% recall on the Greptile benchmark.
Context-aware AI reviewer for Pull Requests. Instant summary, line-by-line comments, title generation and more
A local code review tool designed for the coding agent workflow
```bash
npm install -g diffx-cli
```OrcaCode Review — the open code review harness. Multi-model reviews, merge gates, and no markup. Pay only for inference.
RedFlag uses AI to identify high-risk code changes. Run it in batch mode for release candidate testing or in CI pipelines to flag PRs and add reviewers. RedFlag's flexible configuration makes it valuable for any team.
Review behavior, not just diffs. Jev prioritizes human attention; OpenAI explains the changes. Local CLI + agent skill + GitHub extension.
A code auditing skill by the Oriental Hidden Hero team
Two-model debate review of PRs and MRs, plus a babysitter that works the review rounds. Skills for any coding agent.
Intelligent Git commit code reviewer. Supports local (Ollama) and cloud models (DeepSeek, OpenAI, Anthropic) with a rich CLI UI. 本地AI智能 Git 提交代码审查工具。支持使用本地 (Ollama) 及云端模型 (DeepSeek, OpenAI, Anthropic)，提供丰富的 CLI 交互界面。
Use GPT-5.5 Pro (web-only, no API) from inside Claude Code — repomix-pack the relevant code, send to subscription ChatGPT Pro, harvest the review (API cost $0)
Subagent Verification for Claude AI Code Networks 2026
Self-hosted AI code review for GitHub pull requests, built on Cloudflare Workers.
AI-powered CLI for reviewing GitHub pull requests, detecting potential bugs, security risks, regressions, and missing tests, with structured Markdown reports for open-source maintainers.
Multi-axis PR review orchestration for Claude Code — 5 independent review perspectives, symmetric cross-verification, deterministic coverage assertion
Distilled Code Reviewer skills/soul from 32k~ email from the kernel newsletter
Agentic code reviewer for GitHub PRs and GitLab MRs. Multi-step reasoning with autonomous tool orchestration — catches bugs requiring analysis across files.
```bash
# Just run it (zero install, always latest)
npx @mrkaran/hodor <PR_URL>
# Or install globally
npm install -g @mrkaran/hodor
```codexqa: 11 local-first Agent Skills for Cursor, Claude Code, Codex & OpenClaw — change impact analysis, AI code review, defect scan, testcase generation, browser replay & RCA.
Self-Hosted LLM Code Review Agent for GitLab, Forgejo and GitHub. Local LLM support
Evidence-grounded issue triage, pull request review, and release notes for GitHub maintainers. CLI + GitHub Action.
An open-source multi-agent platform for GitLab MR code review, governance, and CI self-healing.
Brutally honest senior-engineer code reviews for Claude Code, Cursor & Windsurf - and your terminal. Scores, evidence-backed issues, usable fixes.
Salesforce DX code review: Apex, LWC, Flows, security (CRUD/FLS), governors, and Bugbot.
Prototype Gitlab Webook that submits code to OpenAI for review
```
https://git.facha.dev/facha/openai-gitlab-pr-review.git
cd ai-code-reviewer
```A practical example of integrating AI code review into a real-world codebase. Built with Next.js and Firebase, this repo highlights structured prompts, reusable review skills, and parallel subagent execution via Claude Code and OpenAI Codex.
AI-powered code quality analysis using MCP to help AI assistants review code more effectively. Analyze git changes for complexity, security issues, and more through structured prompts.
Claude Code skill for reviewing and designing code based on Clean Architecture principles (Dependency Rule, Layer Structure, SOLID)
AI Code Review Helper automates code review with GitHub/GitLab webhooks and LLMs, posting feedback to PRs/MRs and notifying WeCom.
mr-agent: A Node.js service that auto-triggers AI-powered code reviews when receiving Git merge request webhook events. Integrates with GitHub/GitLab/Bitbucket to analyze changes, flag issues, and suggest improvements—streamlining reviews for teams of all sizes.
AI-powered SAST scanner that finds auth bypass, IDOR, and logic bugs Semgrep/CodeQL miss. Free GitHub Action. Supports Python, JS/TS, Go, PHP, Ruby.
The review app for code your agent writes. Native Rust/GPUI desktop app with real Git workflows.
学习 learn-claude-code 后的 PR Review Agent 实践项目：构建 Debate Council 多 Agent 审查、Tool Calling、结构化报告与 AI Judge 评估。
Automated pull requests reviewing and issues triaging with ChatGPT.
```yaml
name: ChatGPT Review
on: [pull_request]
jobs:
chatgpt-review:
name: ChatGPT Review
runs-on: ubuntu-latest
steps:
- uses: feiskyer/ChatGPT-Reviewer@v0
name: ChatGPT Review
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
OPENAI_API_KEY: ${{ secreAI-powered code analysis tool with bug detection, code explanations, and AI-powered suggestions.
A GitHub Action that uses OpenAI ChatGPT to do code review
```yaml
name: OpenAI ChatGPT Code Review
uses: adshao/chatgpt-code-review-action@v0.2
with:
PROGRAMMING_LANGUAGE: 'JavaScript'
REVIEW_COMMENT_PREFIX: 'chatgpt:'
FULL_REVIEW_COMMENT: 'chatgpt'
OPENAI_TOKEN: ${{ secrets.OPENAI_TOKEN }}
GITHUB_TOKEN: ${{ secrets.GH_TOKEN }}
```Multi-model collaborative development platform — Claude + GPT debate, review, and build together
```bash
# Install globally
npm install -g @codemoot/cli
# One-command setup: verifies codex, creates config, runs first review
codemoot start
# Or step by step:
codemoot doctor # check prerequisites
codemoot init # create .cowork.yml
codemoot review src/ # review code wiKnowledge graph for token-efficient code reviews -- semantic search and call-graph resolution across your codebase.
AI-powered PR review for Gitea/Forgejo built on OpenCode. Runs in Gitea Actions, fetches PR diffs, and posts structured reviews. Supports Docker or source install with isolated .opencode-review config.
Use LLM to help you with PR review
```
pip install pr-review-bot
```On-demand, context-aware AI code reviews in your Git workflow or IDE as you code.
Free, AI PR reviewer that runs entirely in GitHub Actions. No hosting required.
An AI code review agent based on the AgentScope Java SDK and a tutorial covering ReAct, tools, and multi-agent collaboration.
Robin is your AI pull request reviewer in the form of a super fast lightweight github action.
Portable fresh-agent QA loop prompt for finding bugs without biased self-checks / 通用自动 QA 循环 SKILL
A robust GitHub Actions workflow that provides AI-powered code reviews using Amp or Claude Code
Local AI pull request reviewer — review GitHub and BitBucket PRs right from your terminal, powered by the Claude Code, Gemini, and Codex CLIs.
Read what your agent wrote before you say LGTM. A fast TUI in Zig for reviewing code with vim motions.
Ghost personas review your code before it ships. Find edge cases, race conditions, and business logic bugs before production.
AI-powered code review CLI with multiple providers (Gemini, Claude, OpenAI). Features 95%+ token reduction via semantic chunking, 7 review types (security/performance/evaluation), multi-language support, interactive fixes, and developer skill assessment.
Context-aware, AI-assisted code review CLI tool that prevents unnecessary AI-generated code, detects duplicated logic, enforces architectural patterns, and ensures high-quality engineering standards
Open source Claude skill for automated GitLab MR reviews. Configurable lint rules, semantic analysis, and inline comment posting for self-hosted GitLab instances.
An AI-managed open-source code review and assistance tool to help developers write better, more efficient, and secure code.
Professional GitHub PR review skill for Claude Code with pending reviews, code suggestions, and user approval workflow
```bash
# Add this marketplace
/plugin marketplace add aidankinzett/claude-git-pr-skill
# Install the plugin
/plugin install github-pr-review
# Verify installation
/plugin list
```A native /slop-review window for Claude Code, Codex CLI, and pi — review the slop before you ship it. Monaco-powered inline comments, agent reads them back. Forked from badlogic/pi-diff-review.
Expert AI code reviews Catch bugs before they hurt Run Scanline in your CLI to find: - race conditions - logical errors - security risks - optimizations
Stop giving away your code. NeonDIFF is your local-first AI PR reviewer for teams that want professional quality code reviews and issue enrichment without handing every diff to a hosted review SaaS.
Conversational code review on macOS: review a PR or local branch together with a codex agent, and ask about any line instead of just reading its findings.
🔍 Agent Verifier is a coding agent skill that verifies code against organizational policies, code quality patterns, security requirements, and framework best practices — before code ships. Works with Claude Code, Cursor, Windsurf, and 30+ agents.
Stop trusting single AI outputs. An anti-hallucination framework with dual-brain arbitration, multi-model audit, and evidence chains for LLM quality control.
🤖 AI code quality gate for AI-generated code. Detects hallucinated packages, phantom dependencies, stale APIs, and more. MCP Server + CLI + CI/CD Action.
Context-Driven Development plugin for Claude Code, Cursor, GitHub Copilot, Gemini
Local, live code review for the AI age - read what the agent wrote and send feedback straight back
Local-first multi-runtime AI cockpit (CLI + desktop + MCP). War-rooms, review receipts, open multi-model coding benchmarks. MIT.
Review diffs in a terminal UI, leave inline comments, and let coding agents fix them
Code review agent for Claude Code. One .md file — 14 review categories, 4 severity levels, auto-fix, OWASP security checks. Zero dependencies.
Checks AI-generated code changes before merge: scope, validation, risk, review evidence, and optional Verity receipts.
AI-powered git pre-commit hook for automated code review with customizable rules.
Design system guardian for code. Anchors to design system, catches drift in React/Vue/Tailwind, reviews PRs automatically. Free for open source.
A SKILL that turns a code diff into an evidence-backed ledger of hidden assumptions, failure modes, and falsification tests.
Review-only AI code review for GitHub pull requests. Cross-check PRs with model-agnostic reviewers that can comment inline but cannot approve, merge, or close.
8 AI agents grade your entire repository across 6 dimensions — signed, verifiable A+ to F audit reports in minutes
MCP server for AI-to-AI collaboration — bridge Claude with Gemini, Codex, and other LLMs for code review, second opinions, and plan debate
Context Aware Security Gate for GitHub Actions. Validates deployments using Hybrid Analysis (Deterministic Rules + LLM Reasoning).
One-command AI code quality pipeline: review, security, tests, performance, simplify & cleanup. Health scores, CI gate. For Cursor, Claude Code, Qoder, Trae.
使用 TypeSafe Jev 审查 Skill 与 MCP 可疑行为 | Review Agent Skills and MCP code with Jev, static evidence, and explicit coverage gaps
Evidence-based review skill for diff, project, and artifact review across code, readiness, architecture, requirements, and risk.
🚦 Senior-QA review skill for Claude Code: runs a regression hunter + a change reviewer in parallel, then gives one read-only SIGN OFF / DO NOT SHIP verdict with file:line evidence.
Filter AI slop before it hits your codebase. A structured human review workflow for AI-generated code.
AI writes code faster than you can review it. HighStakes reads your codebase, scores every file by blast radius, and tells you where human review matters and where AI can handle the rest.
Claude Code workflow to grade your project's maintainability by Human and LLM
Full-spectrum design patterns skill: Detect anti-patterns, diagnose architectural debt, generate production-ready refactored code. GoF, SOLID, DDD, CQRS, Microservices.
Local LLM code review CLI for git diffs — review staged/unstaged changes, any git diff range, or a GitHub PR, right in your terminal.
20 AI agents in your git workflow. Secrets, injection flaws and unsafe migrations are blocked at commit; performance, coverage and the rest are reviewed at push. Research → Plan → Implement workflow with GO/NO-GO gates before you write a line. Generates production-ready code, product wikis and codebase maps on demand. Any language. Free.
Terminal code review with vi-like bindings. Live diffs, comments the agent answers, PR review, MCP built in.
Local web UI to review AI-generated code changes and generate structured feedback prompts
Agentic code review CLI — LangGraph orchestrates 13 deterministic static analyzers (oxlint/bandit/clippy/spotbugs/golangci/cppcheck) + LLM reasoning across 23 languages including industrial PLC. Grounding validator kills hallucinated paths. RAG over your guidelines. dedup --fix actually edits files with multi-step undo (no git required).
⚡ Preflight — AI code review before you open the PR Multi-agent system that gives instant, senior-level feedback on git diffs.
OBJECTION! A skill and plugin for AI coding agents (Claude Code, Codex, Cursor, Gemini): accusers, a defender and a judge review the PR, and a gate blocks it until APPROVED.
Dual-AI automated code review: Codex finds issues, Claude fixes them — loop until clean.
AI-powered code review platform that automatically analyzes GitHub PRs for security vulnerabilities, bugs, and complexity issues — posts inline comments like a real senior developer.
🦨 Stop shipping smelly code. 120+ named code smells with before/after examples across C, C++, C#, Java, Python, Go, Rust, JavaScript, TypeScript & SQL — an AI agent skill that improves code quality during both development and code review.
Skeptical ship/no-ship adversarial code review of a git diff or branch — runs against any LLM (API or local CLI agent)
A CLI that orchestrates parallel AI coding agents to reciew, fix and ship code changes
| Tool | Stars | Score |
|---|---|---|
| open-code-review | ★ 46.5k | 82 |
| codex-plugin-cc | ★ 34.1k | 71 |
| code-review-graph | ★ 31.6k | 76 |
| pr-agent | ★ 13.3k | 77 |
| DeepAudit | ★ 7.1k | 65 |
| claude-code-security-review | ★ 6.4k | 74 |
| shippie | ★ 2.5k | 68 |
| mantis | ★ 2.4k | 78 |
| code-review-skill | ★ 1.9k | 75 |
| sourcery | ★ 1.9k | 61 |
| px0 | ★ 1.8k | 66 |
| roborev | ★ 1.7k | 66 |
| openreview | ★ 1.7k | 58 |
| brooks-lint | ★ 1.5k | 76 |
| git-lrc | ★ 1.5k | 67 |
| kodus-ai | ★ 1.5k | 64 |
| sashiko | ★ 1.3k | 66 |
| gentleman-guardian-angel | ★ 1.2k | 59 |
| guard-skills | ★ 1.2k | 83 |
| ai-codereviewer | ★ 1.0k | 58 |
Yes. openai/codex-plugin-cc lets Claude Code hand a review to Codex, and awesome-skills/code-review-skill is a review skill covering React, Vue, Rust, TypeScript and more. For reviews by several models that check each other: chorus-codes/chorus. They are listed on this page under Agent skills and Multi-agent.
The-PR-Agent/pr-agent reviews pull requests on GitHub, GitLab and others; alibaba/open-code-review is a code review engine used at Alibaba's scale; kodustech/kodus-ai is self-hosted and works with your own LLM. They are listed on this page under PR bots & CI, Local & CLI and Platforms.
The top code review tools in 2026 are open-code-review, codex-plugin-cc, code-review-graph. Agent Skills Hub ranks 185 options by GitHub stars, quality score (6 dimensions including completeness, examples, and agent readiness), and recent activity. The list is rebuilt every 8 hours from live GitHub data.
open-code-review (46.5k stars) is the most adopted choice for general code review workflows, written in Go. codex-plugin-cc (34.1k stars) is a strong alternative and uses JavaScript instead. Pick by your existing stack: match the language and runtime your team already uses to minimize integration cost. If unsure, start with open-code-review — it has the deepest community and the most examples online.
Avoid pre-built code review tools when (1) your use case requires deep customization that the tool's plugin system doesn't support, (2) you have strict compliance requirements that ban third-party dependencies, (3) the tool's maintenance is inactive (last commit >6 months ago), or (4) your data volume is small enough that a 50-line custom script is cheaper than learning the tool. For most production workflows above 100 requests/day, the time savings from a maintained tool outweigh the customization loss.
Code Review focuses specifically on claude code review skills and ai code review tools for claude code and codex: pr review bots, local cli reviewers, security review and multi-agent review. every one is security-graded before you install it. Claude Code Hooks, Subagents & Statusline is a related but distinct category — see https://agentskillshub.top/best/claude-code-hooks/ for those tools. The two often appear in the same agent pipeline but solve different problems: choose code review when your primary goal is the specific task, and claude code hooks, subagents & statusline when the workflow is broader.
For most teams, yes. open-code-review has 46.5k stars worth of community testing, handles edge cases you haven't thought of, and ships with documentation. Build your own only when (1) your requirements are deeply non-standard, (2) you have a security/compliance reason to avoid OSS dependencies, or (3) the maintenance burden is small enough (<200 lines of code) that you'll save time long-term. The break-even point is usually around 2-3 weeks of dev time saved.
Most code review tools listed are open source under permissive licenses (MIT, Apache 2.0). A handful offer paid managed/cloud versions on top of free self-hosted core. Always check the LICENSE file on each tool's GitHub repository before commercial use — some use AGPL or non-commercial restrictions that may not fit your deployment model.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: