{"categories":[{"label":"Artificial Intelligence","url":"https://skillfed.io/packages/category/scientific-engineering-artificial-intelligence"}],"enrichment":{"capability":"An AI agent that solves software engineering tasks by running bash commands in sandboxed environments, controlled by language models via litellm.","skillfed_tags":["ai-agent","code-generation","llm-automation"],"use_cases":["Automated code generation and bug fixing in software repositories using language model reasoning.","Benchmarking language model coding capabilities on SWE-bench or similar evaluation frameworks.","Building custom AI agents for software tasks by extending the simple agent scaffold with domain-specific prompts.","Running AI-driven code analysis and refactoring workflows in isolated sandbox environments.","Fine-tuning language models on agent trajectories without overfitting to complex agent infrastructure."],"what_it_does":"mini-swe-agent is a stripped-down AI coding agent that uses language models to solve software engineering tasks by executing bash commands in isolated environments. Unlike more complex agent frameworks, it has no custom tools beyond bash and maintains a linear message history, making it transparent for debugging and fine-tuning. It supports multiple deployment modes (local, Docker, Singularity, bubblewrap) and works with any model accessible through litellm, openrouter, or portkey.\n\nThe package is designed for developers who want a hackable baseline rather than a black-box system. It includes a CLI tool for interactive use, batch inference capabilities, and Python bindings for programmatic control. With 14 runtime dependencies (pyyaml, requests, jinja2, pydantic, litellm, tenacity, rich, python-dotenv, typer, platformdirs, textual, prompt_toolkit, datasets, openai), it provides a complete but minimal agent stack.","worth_installing":"Yes, if you need an AI agent for code tasks and want transparency and simplicity over feature richness. The low install friction, active maintenance, permissive MIT license, and zero known vulnerabilities make it a safe choice. The alpha status and reliance on external LLM APIs (not included) are minor caveats. Best suited for developers, researchers, and teams already comfortable with language model APIs."},"id":"mini-swe-agent","links":{"html":"https://skillfed.io/packages/mini-swe-agent","md":"https://skillfed.io/packages/mini-swe-agent.md","pypi":"https://pypi.org/project/mini-swe-agent/"},"maintenance":{"status":"active"},"meta":{"latest_release":"2026-07-23","license_spdx":null,"license_treatment":"permissive","name":"mini-swe-agent","python_support":"supports_current","summary":"Mini SWE Agent - A simple AI software engineering agent"},"popularity":{"monthly_downloads":15501428,"position":1184,"tier":"top_5000"},"security":{"n_vulnerabilities":0},"version":"2.4.6"}