"Harness^2: Recursive Agent Harnessing for an Open World"
摘要
Google 提出 Harness²，用两层递归自我改进让智能体框架适配开放世界任务：针对当前任务提出、测试并组合对提示、工具与工作流的修改，再把实验经验存储下来用于改进未来的框架调整。
摘要由模型根据原文撰写，以原文为准。
正文 · 原文
"Harness^2: Recursive Agent Harnessing for an Open World"
Harness optimization would usually struggle to adapt to new tasks without ground-truth feedback, and improvements made for one task rarely transfer to others too.
So Google's Harness² introduces two levels of recursive self-improvement: it proposes, tests, and combines changes to an agent's prompts, tools, and workflows for the current task, then stores lessons from those experiments to improve future harness modifications.
Across GLM-5 and Gemini 3.5 Flash, it improves performance by up to 11.5% on WorkBuddy-Bench, 4.5% on JobBench, and 11.8% on Legal Agent Bench, without updating model weights.
来源：alphaXiv · x.com
引用原文
askalphaxiv. "Harness^2: Recursive Agent Harnessing for an Open World"[EB/OL]. (2026-10-21)[访问日期]. https://x.com/askalphaxiv/status/2112863284147724678.