SWE-agent
SWE-agent lets your language model autonomously use tools to fix GitHub issues, find cybersecurity vulnerabilities, or perform custom coding tasks. Built by Princeton and Stanford researchers, it achieves state-of-the-art results on
Visit SWE-agent/SWE-agentOverview
SWE-agent enables your language model of choice to autonomously use tools to fix issues in real GitHub repositories, find cybersecurity vulnerabilities, or perform any custom task. It is built and maintained by researchers from Princeton University and Stanford University.
Key Features
- State of the art on SWE-bench among open-source projects
- Free-flowing and generalizable: leaves maximal agency to the language model
- Configurable and fully documented: governed by a single YAML file
- Made for research: simple and hackable by design
- Supports multimodal input for processing images from GitHub issues with vision-capable models
Use Cases
- Fixing issues in real GitHub repositories
- Finding cybersecurity vulnerabilities (offensive cybersecurity / CTF challenges)
- Performing custom coding tasks and competitive coding challenges
Getting Started
- Install from source or try it in your browser via GitHub Codespaces. See the documentation for installation and hello world tutorials.
Deployment & Requirements
- Requires a language model API key (e.g., GPT-4o or Claude Sonnet 4) and a Python environment. Configuration is done via a single YAML file.
Before You Adopt
- License: MIT. Review its terms before using, modifying, or distributing the project.
- SWE-agent has been superseded by mini-swe-agent and is now in maintenance-only mode.
