自我优化框架：循环工程实现自主改进

self-harness: harnesses that improve themselves

spoiler alert, it's all loop engineering!

1. weakness mining - run agent and observe failures 2. propose improvements to the harness 3. confirm improvements are beneficial and don't cause other regressions

and then you loop!

the paper: https://arxiv.org/pdf/2606.09498 powered by deepagents: https://github.com/langchain-ai/deepagents