Skip to content
Flows
Tier 4 · Open-source coding agents Coding agents · fetched 2026-07-02

SWE-agent

The Princeton research agent that introduced the Agent-Computer Interface (ACI) concept: carefully designed commands, file viewers, and feedback formats that dramatically improve LLM performance on real GitHub issues (SWE-bench).

View source on GitHub

Key takeaways

  • 01

    ACI design (commands + viewers + feedback) is a measurable performance lever

  • 02

    Constrained, purpose-built tools beat raw shell access

  • 03

    Benchmark-driven iteration (SWE-bench) keeps harness claims honest

Flows built on this research