SWE-agent

SWE-agent (Princeton) frames software engineering as an agent interacting with a controlled interface (ACI) optimized for language models resolving GitHub issues. It is influential research/OSS for agent–repo interaction design. Use as evaluation and design reference for coding agents; not by itself an enterprise OS runtime. SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice. It can also be employed for offensive cybersecurity or competitive coding challenges. [NeurIPS 2024] [!warning] Most of our current development effort is on mini-swe-agent, which has superseded SWE-agent. It matches the performance performance of SWE-agent, while being much simpler. See the FAQ for more details about the differences. Our general recommendation is to use mini-SWE-agent instead of SWE-agent going forward. SWE-agent enables your language model.