SWE-agent autonomously resolves GitHub issues and software engineering tasks by letting language models use tools in real repositories. Current development has shifted to mini-SWE-agent, which the project recommends for general use.
Project overview
SWE-agent autonomously resolves GitHub issues and software engineering tasks by letting language models use tools in real repositories. Current development has shifted to mini-SWE-agent, which the project recommends for general use.
Project type
AI Agent · AI Coding
Use cases
Coding & Development
Deployment
Refer to project documentation
License
MIT
Best for
Researchers and developers needing a hackable, YAML-configurable system for autonomous software engineering tasks.
Users looking to explore autonomous issue resolution or offensive cybersecurity modes using language models.
Key capabilities
Takes a GitHub issue and tries to automatically fix it using the language model of choice.
Can be employed to perform any custom task, such as competitive coding challenges.
A mode for solving offensive cybersecurity (capture the flag) challenges.
Limitations and risks
Most current development effort is on mini-swe-agent. The general recommendation is to use mini-SWE-agent.
Setup difficulty, GPU requirements, cost dependencies, and data boundary constraints are not documented.
Getting started
The README provides links to installation instructions and GitHub Codespaces, but specific setup steps are not detailed.
Evidence and sources
GitHub project description: SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice.
README: SWE-agent enables your language model of choice (e.g. GPT-4o or Claude Sonnet 4) to autonomously use tools to fix issues in real GitHub repositories, find cybersecurity vulnerabil…
README: Most of our current development effort is on [mini-swe-agent](https://github.com/SWE-agent/mini-swe-agent/), > which has superseded SWE-agent.
README: ✅ Made for research: Simple & hackable by design
README: ✅ Configurable & fully documented: Governed by a single yaml file