program-repair

Kodezi Chronos is a debugging-first language model that achieves state-of-the-art results on SWE-bench Lite (80.33%) and 67% real-world fix accuracy, over six times better than GPT-4. Built with Adaptive Graph-Guided Retrieval and Persistent Debug Memory. Model available Q1 2026 via Kodezi OS.

Software development bots for Github. Join the bot revolution! 🌟🤖🌟💞

Automatic program repair for Java with generate-and-validate techniques ✌️✌️: jGenProg (2014) - jMutRepair (2016) - jKali (2016) - DeepRepair (2017) - Cardumen (2018) - 3sfix (2018)

Automatic program repair system for Java based on dynamic analysis and code synthesis with SMT. Also contains the code of Dynamoth.

Semantic program repair system for C programs

Automatic repair system for static analysis warnings from SonarQube's SonarJava, TDSC 2022 http://arxiv.org/pdf/2103.12033

✅SRepair: Powerful LLM-based Program Repairer with $0.029/Fixed Bug

A High-performance, High-extensibility and High-portability APR System

For our ISSTA23 paper "How Effective are Neural Networks for Fixing Security Vulnerabilities?" by Yi Wu, Nan Jiang, Hung Viet Pham, Thibaud Lutellier, Jordan Davis, Lin Tan, Petr Babkin, and Sameena Shah.

This is an automated transformation inference tool that leverages a big code corpus to guide the abstraction of transformation patterns.

Open-science code of the program repair tool described in "Dynamic Patch Generation for Null Pointer Exceptions Using Metaprogramming" (SANER'17)

A dataset of Java bugs for automatic repair, derived from the C bugs of IntroClass

基于 AI Agent 服务自动化修复系统：Agent 自动读取错误日志，定位 Bug，生成补丁，运行测试，提交 PR，并通知开发者 Review。AI-powered auto-fix agent for web services: analyzes logs, patches code, runs tests, and creates pull requests automatically.

A search-based automated program-repair tool for Java with LLM integration.

Refactoring-Aware Patch Integration Across Structurally Divergent Java Forks

Experiments in code security using Python, JavaScript, Java, and Go.

Four-arm study of Java bug fixing: execution inverts the CodeBLEU ranking. RAG-prompted Qwen fixes 35.8% of 201 real bugs; LoRA tops CodeBLEU at 0.854 but fixes fewer.

51 Java regression bugs from 28 open-source projects (extracted from BugSwarm & Bears benchmarks)

Java implementation of Custom GenProg algorithm for automated program repair using genetic programming