Open sourcePython · ≈8.9k stars · MIT
mini-swe-agent
The ~100-line coding agent that scores >74% on SWE-bench Verified.
SWE-agent/mini-swe-agent
Maintained by SWE-agent (Princeton & Stanford)
mini-swe-agent is a deliberately minimal software-engineering agent from the Princeton and Stanford team behind SWE-bench and SWE-agent. The agent class is about 100 lines of Python, with small environment, model and run modules, and the project reports scores above 74% on SWE-bench Verified. It runs in local, Docker/Podman, Singularity/Apptainer and bubblewrap environments and supports any model via litellm, OpenRouter or Portkey. The README lists users including Meta, NVIDIA, IBM and Anyscale and says it powers Ramp's SWE-Bench. Now at v2.
In research (18)
- BenchMeUp: Building Software Engineering Benchmarks from Natural-Language SpecificationsVerdict: Useful
Oct 27, 2026 · BenchMeUp's search and construction agents use a customised mini-swe-agent scaffold.
Oct 27, 2026 · One of the five coding harnesses compared (59.8%).
- Skill Dreaming: Evolving Repository Skills through Pull Request StreamsVerdict: Incremental
Oct 26, 2026 · The base coding agent into which learned skills are injected.
- Do Coding Agents Still Need a Debugger? Frontier-Model-Based Fault Localization at the Accuracy CeilingVerdict: Incremental
Oct 26, 2026 · A bash-only agent loop with no debugger; this study's finding that strong models localise faults without stepping tools supports that minimal design.
- CIVER: Causal Intervention Verification for Reliable Code ReviewVerdict: Incremental
Oct 26, 2026 · The 'fixed base mini agent' used in the downstream repair experiment.
- V$^3$AST-Bench: Verifiable, Valid, and Versatile Safety Evaluation for Coding AgentsVerdict: Important
Oct 26, 2026 · Lightweight baseline harness evaluated.
Showing the 6 newest of 18 papers.
Neighbours on the workbench
MDX · ≈307k stars · MIT
obra/superpowers
An agentic skills framework and software methodology that works.
Maintained by Jesse Vincent (obra)
Shell · ≈306k stars · MIT
mattpocock/skills
Small, composable agent skills for real engineering work.
Maintained by Matt Pocock (AI Hero)
JavaScript · ≈292k stars · MIT
affaan-m/ECC
Agent harness performance optimization system: skills, instincts, memory and security for coding agents.
Maintained by Affaan Mustafa
Connections
Relations from the Harness Hub knowledge graph, as charted in the State of Harness reports.
Maintained by · 1
- SWE-agentOrganisation
Similar entries · 2
- AutoCodeRoverProject
- SWE-agentProject
Mentioned in · 1
- SWE-VERIFIER: Selective Verification for Cost-Efficient Coding AgentsPaper · 2026-10-21
Sources
Verified Oct 29, 2026