SWE-smith
ActiveDescription
Scaling data for SWE-agents (NeurIPS 2025 D&B Spotlight). A toolkit for automatically generating large-scale training datasets for software engineering agents.
Related Projects
Multi-SWE-bench
331 · Python
A multilingual benchmark for issue resolving. Extends SWE-bench to multiple programming languages for evaluating AI agent capabilities across diverse codebases.
benchmarkswe-benchmultilingual +3
SWE-AF
707 · Python
SWE-AF is an autonomous software engineering fleet platform using a multi-agent factory architecture. It orchestrates planner, coder, reviewer, and verifier agents to automate the full software engineering lifecycle from issue analysis to code fix, scoring 95/100 on benchmarks.
swe-agentmulti-agentautonomous-engineering +3
AutoCodeRover
3.1k · Python
AutoCodeRover is a project structure-aware autonomous software engineer agent that achieves automated program repair and issue resolution by understanding the overall codebase architecture.
codingpythonagent +2
DeepCode
15.2k · Python
DeepCode is an open agentic coding platform supporting Paper2Code, Text2Web, and Text2Backend, leveraging agent technology for automated software development workflows.
codingpythonllm +2