mini-swe-agent
Fewer than 5 people recommend it so far, so there's no written verdict. Here is what commenters said.
Why people recommend it
“normalizing the harness to mini-swe-agent”
“There's a ~100LoC harness called mini-swe-agent [1] that does just that.”
Show 3 more
“Look at https://github.com/SWE-agent/mini-swe-agent for proof”
“Check out [1] a 100 LoC "LLM in a loop with just terminal access", it is now above last year's heavily harnessed SotA.”
Caveats and dissent
No one in these comments warned against it.
From the makers not counted in the ranking
“I wrote https://github.com/swe-agent/mini-swe-agent/ , which is probably the most minimal agent out there”
“I think it's a great way to dive into the agent world”
Threads where it came up
- An empirical study of harness design for coding agents 2 mentions
- How to Code Claude Code in 200 Lines of Code 2 mentions
- DeepSWE: A contamination-free benchmark for long-horizon coding agents 1 mention
- Parametric CAD Bench 1 mention
- Scaling long-running autonomous coding 1 mention
- Claude Cowork Exfiltrates Files 1 mention
Data: mini-swe-agent.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 8
“190 lines with comments included so close enough”
“normalizing the harness to mini-swe-agent”
“why opus 4.7 is with mini-swe-agent, not claude code”
“There's a ~100LoC harness called mini-swe-agent [1] that does just that.”
“Check out mini-swe-agent.”
“Look at https://github.com/SWE-agent/mini-swe-agent for proof”
“Check out [1] a 100 LoC "LLM in a loop with just terminal access", it is now above last year's heavily harnessed SotA.”
“Check out the exercise from the swe-agent people who released a mini agent that's "terminal in a loop"”