dsadaq/miniswe
mini-swe-agent prompt search notes 这套仓库原本就分成两层能力,但之前入口不清楚,很容易让人误以为“agent 会在单次运行里自己联网、自己改 prompt”。 运行时联网 src/minisweagent/config/benchmarks/swebench.yaml 会把 mswea-web-search 和 mswea-web-fetch 装进容器,并且 prompt 里会提示模型可以用它们查公开文档。 Prompt 迭代 scripts/search_system_prompt.py 会跑一个离线的 prompt policy search。它会: 生成 prompt override 跑一轮 SWE-bench canary 分析失败轨迹 把失败模式反馈到下一轮 prompt 搜索 它不是 agent 在同一次任务里“边做边改 system prompt”,而是外部搜索脚本驱动的多轮评测闭环。 现在可直接从主脚本开启 prompt search… See the full description on the dataset page: https://huggingface.co/datasets/dsadaq/miniswe.
| Author / organization | dsadaq |
|---|---|
| Downloads | 789 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 26 May 2026 |
| Last updated at source | 26 May 2026 |
Popularity, last 2 days
789 now (+49 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.