DeepAudit：人人拥有的 AI 黑客战队，让漏洞挖掘触手可及。国内首个开源的代码漏洞挖掘多智能体系统。小白一键部署运行，自主协作审计 + 自动化沙箱 PoC 验证。支持 Ollama 私有部署 ，一键生成报告。支持中转站。​让安全不再昂贵，让审计不再复杂。

AI-powered code analysis tool with bug detection, code explanations, and AI-powered suggestions.

An automatic API misuse checker for C programs!

Logic-first AI code review via semi-formal execution tracing (Premises → Trace → Divergence → Trigger → Remedy). Catches behavioral bugs, type-contract breaches & async hazards that linters miss. Six skills · Claude Code · Codex CLI · Gemini CLI.

Python package to predict bugs using the complexity of code changes

A graph based bug classifier using the dgl library and DeepBugs dataset

Generate more realistic mutations with contextual mutants

AI-powered GitHub repository analyzer that ingests full codebases, enables codebase-wide chat, detects bugs, and suggests optimizations using RAG (LangChain + FAISS + OpenRouter) with a Streamlit UI.

An attempt at detecting SStuBs using a pre-trained transformer and repairing them with a seq2seq model.

Autonomous QA MCP that tests web and macOS apps like a real testing engineer—and verifies every bug it reports.

The Lighthouse for Android Apps — automated QA audit with Production Readiness Score. One command, one report. Open source.

Self-supervising TLA+ formal verification loop — an LLM agent iteratively writes specs and fixes bugs, using the TLC model checker as an incorruptible evaluator. Inspired by Karpathy's autoresearch.

formal analysis for the codebases AI builds

Reimplementation of Anthropic's Natural Language Autoencoder on Qwen2.5-0.5B, with GRPO RL training, layer ablation, and a code-model extension probing bug-semantic content in Qwen2.5-Coder.

The AI Developer's Assistant aims to enhance developer productivity by analyzing code snippets, identifying bugs, suggesting optimizations, generating unit tests, and debugging log files.

An honest, open benchmark of AI code-review tools against projects with a known set of planted bugs — measuring recall, false positives, speed, and cost.

AI-powered visual bug hunter for GUI apps. Automatically detects UI anomalies, locates source code, and generates minimal diffs to fix bugs. Designed for Vibe Coding users who need reliable visual verification.

FailMapper: Failure-Scenario-Guided Unit Test Generation using Monte Carlo Tree Search (MCTS) and LLMs. Detects 233% more bugs than baselines on Defects4J. Covers 9 failure scenarios for automated bug detection in Java programs. [ASE 2025]

AI Code Review Memory - learns from your team's bug history and warns when similar patterns appear