awesome-ai-code-review

Awesome list of AI Code Review agents, copilots, and research papers.
https://github.com/kodustech/awesome-ai-code-review

Last synced: about 11 hours ago
JSON representation

Automated PR Agents
- Cursor Bugbot - powered PR review that runs automatically to catch real bugs and security issues with a low false-positive rate.
- Revieko - Repo-specific architecture drift detection in pull requests (structural risk + hotspots).
- Git AutoReview
- AI Change Passport - GitHub Action and CLI that creates signed provenance reports for AI-assisted pull requests, helping reviewers see sensitive files, dependency edits, workflow changes, and test coverage signals.
- Conclave AI - A council of three models (Claude, GPT-5, Gemini) reviews each PR and, with a PRD attached, flags scope drift a diff-only review misses. Includes an autofix worker; self-hostable (FSL).
- Ejentum agent-teams (adversarial-code-review) - agent code-review team built on the Ejentum Reasoning Harness. Reviewer agents call `harness_code` and `harness_anti_deception` as agentic tools during the review loop; each call returns a structured scaffold (named failure pattern, executable procedure, suppression vectors, falsification test) the agent reads internally before issuing review feedback. Catches false-confident approvals and sunk-cost framings that single-agent reviewers miss.
- Gito - Open-source AI code reviewer for GitHub and GitLab. Runs locally via CLI or in CI/CD pipelines, and works with any LLM provider — including self-hosted models (Ollama, LM Studio, vLLM) for fully private reviews.
- Ivy Tendril - Open-source AI coding orchestrator with automated verification gates (build, test, lint, format, AI review) that enforce code quality before any PR reaches human review. Orchestrates Claude Code, Codex, Antigravity, Copilot, and OpenCode.
- Kodus
- CodeRabbit - by-line feedback on pull requests and generates summaries of changes. Features a chat interface within the PR to discuss the feedback with the AI.
- Git AutoReview
- Zenable - commit, and PR review to catch bugs and security issues in AI-generated code in real-time.
- Polarity - The First AI QA Engineer which does code review, testing, and long running agent tasks. Understands your entire codebase and code quality, zero fluff comments.
- Swarm Orchestrator - GitHub Action and CLI that audits AI-generated pull requests for eleven cheat patterns (relaxed tests, swallowed errors, fake renames, and more). Advisory by default, with an opt-in merge gate.
- heygrc
Benchmarks
- Code Review Benchmark - powered code review tasks.
- SWE-bench - Evaluation framework for language models on real-world software engineering issues.
- HumanEval - OpenAI's dataset for evaluating code generation capabilities.
- Code Review Benchmark - powered code review tasks.
- ReviewBenchLite - Comprehensive evaluation of code review agents on top 100 git repos.
- ShopPay Audit Benchmark - Spec-grounded benchmark for testing whether AI code review/coding agents catch business-logic defects in payment, wallet, webhook, and authorization flows.
CLI & Local Workflows
- Aider - AI pair programming in your terminal.
- OpenCommit - Generates semantic git commit messages automatically.
- prpack - Zero-dependency Node CLI that packs a pull request (diff plus the full post-change content of every touched file) into a single markdown file optimized for LLM code review. Companion [GitHub Action](https://github.com/Lucas2944/prpack-action) and [browser demo](https://lucas2944.github.io/prpack-demo/) available.
- LegacyDoc AI - VS Code extension that generates AI code audit reports, architecture maps, and review-ready context packs from a workspace before cleanup or code review.
- Signum - Multi-model code review pipeline that dispatches diffs to Claude, Codex, and Gemini as independent reviewers with adversarial isolation, then bundles findings into a tamper-evident proofpack.
- Mentat - Coordinate edits across multiple files using command line.
- claude-code-multi-model-review - Claude Code skill that invokes multiple AI models (DeepSeek, Doubao, Qwen, OpenAI) in parallel to review git diffs with cross-validation across model outputs.
- h5i - Rust CLI that runs several coding agents (Claude Code, Codex) on the same task in isolated git-worktree sandboxes, has them peer-review each other, then a neutral verifier replays each candidate, runs the tests itself, and merges the one that passes. Run metadata (prompts, models, commands, logs, verdict) is versioned in the repo under refs/h5i/*.
IDE Assistants & Copilots
- GitHub Copilot - The standard AI pair programmer for autocomplete, chat, and inline edits.
- Claude Code - Claude's coding agent for terminal, IDE, and web workflows that can manage large codebases and implement changes.
- OpenAI Codex - OpenAI's coding agent that can read, modify, and run code, available as a VS Code extension with optional cloud delegation.
- Cline - Autonomous IDE agent that can create/edit files, run commands, and use the browser with user approval.
- SimpleReview - Browser extension for in-browser visual code review. Hover any element on a live website, click "Fix it" to get AI-powered fix suggestions in a side panel.
- Google Antigravity - Agent-first IDE with tab autocomplete, natural language commands, and cross-surface agents across editor, terminal, and browser.
- Kilo Code - Open-source agentic engineering platform with IDE/CLI support, tab autocomplete, and multi-agent orchestration.
- OpenCode - Open-source coding agent for terminal, IDE, or desktop with multi-session workflows and broad model support.
- SimpleReview - Browser extension for in-browser visual code review. Hover any element on a live website, click "Fix it" to get AI-powered fix suggestions in a side panel.
Key Research & Papers
- AI-Assisted Assessment of Coding Practices in Modern Code Review - based system for enforcing coding best practices in code review.
- Towards Practical Defect-Focused Automated Code Review
- Automated Code Review Using Large Language Models at Ericsson: An Experience Report
- DeputyDev -- AI Powered Developer Assistant: Breaking the Code Review Logjam through Contextual AI to Boost Developer Productivity
- CRScore++: Reinforcement Learning with Verifiable Tool and AI Feedback for Code Review
- Leveraging Reviewer Experience in Code Review Comment Generation
- A Survey on Machine Learning Techniques for Source Code Analysis
- BitsAI-CR: Automated Code Review via LLM in Practice - based automated code review in large-scale environments.
- AutoCodeRover: Autonomous Program Improvement
- Prompting and Fine-tuning Large Language Models for Automated Code Review Comment Generation - tuning LLMs with QLoRA for generating accurate code review comments.
- Fine-Tuning Large Language Models to Improve Accuracy and Comprehensibility of Automated Code Review
- Lessons from Building Static Analysis Tools at Google - positive rates are crucial (validating the need for specialized agents).
Security & Static Analysis AI
- Snyk DeepCode - AI-powered engine to find security flaws faster than traditional static analysis.

Programming Languages

JavaScript 3 Python 2 Rust 1 HTML 1 C# 1 Shell 1 TypeScript 1

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Awesome

awesome-ai-code-review

Automated PR Agents

Benchmarks

CLI & Local Workflows

IDE Assistants & Copilots

Key Research & Papers

Security & Static Analysis AI