The mini-swe-agent is a minimalist AI software engineering agent designed to solve GitHub issues and assist with command-line tasks, offering a simple and performant solution for developers.
Source: Description per README View on GitHub →This project is gaining attention due to its minimalist approach, which provides a simple and efficient tool for developers. It addresses the pain point of complex and resource-intensive AI agents by offering a lightweight solution that still achieves high performance on benchmarks like SWE-bench. The project stands out for its simplicity, ease of use, and compatibility with various environments and models.
Source: Synthesis of README and project traitsThe agent is designed to be simple and easy to understand, with minimal configuration and dependencies, making it accessible to a wide range of users.
Source: READMEThe agent scores over 74% on the SWE-bench benchmark, demonstrating its effectiveness in software engineering tasks.
Source: READMESupports local environments, Docker, Podman, Singularity, and more, allowing for flexible deployment options.
Source: READMEWorks with various models via litellm, openrouter, portkey, and more, providing flexibility in choosing the right model for the task.
Source: READMEThe architecture is modular, with clear separation between the agent class, environment, model, and run script. It uses subprocess.run for executing actions, ensuring independence and ease of sandboxing. The linear history of the agent simplifies debugging and fine-tuning.
Source: Code tree + dependency filesCenter: project; inner ring: core feature modules; outer ring: key dependencies. Auto-generated from core_features and tech_stack.key_deps.
pyyamlrequestsjinja2pydanticlitellmtenacityrichpython-dotenvtyperplatformdirstextualprompt_toolkitdatasetsopenaiThe project is suitable for developers who need a simple and efficient AI agent for GitHub issue resolution, command-line assistance, or software engineering tasks. It is particularly useful for those working on local environments or looking for a lightweight solution for benchmarking and testing AI agents.
Source: READMEv2.2.8 (2026-03-24): Fixed issues with the litellm dependency and other minor fixes. v2.2.7 (2026-03-12): Fixed issues with the ContreeEnvironment. v2.2.6 (2026-03-02): Fixed a bug in exception handling with the SwerexModalEnvironment. v2.2.5 (2026-02-27): Bugfix release for FormatError issues and other rare issues. v2.2.4 (2026-02-23): Fixed a bug causing an exception about missing default model name.
Source: GitHub ReleasesThe mini-swe-agent is a promising project for developers seeking a simple, high-performance AI agent for software engineering tasks. Its minimalist design and cross-platform compatibility make it a valuable tool for both individual developers and organizations looking to integrate AI into their workflows.