SkillOpt：将智能体技能作为可训练参数，实现可靠性能提升

AI 摘要

微软研究院发布SkillOpt，一种将智能体技能文件视为可训练参数的新方法，在不更新模型权重的情况下，通过前向-反向-更新循环优化技能。在六个基准、七个模型和三种执行模式下，SkillOpt在全部52个评估单元中取得最佳或并列最佳结果，显著提升性能，且优化后的技能可跨模型和任务迁移。

正文节选

At a glance - AI agents often fail because their instructions, or skills, are manually modified with no guarantee of improvement. SkillOpt turns skill editing into a training process, making agent behavior more reliable without changing model weights. - SkillOpt treats an agent skill file as a trainable parameter outside a frozen target model, turning skill writing from one-shot prompting into a controlled optimization process. - Across six benchmarks, seven target models, and three execution mo

发布时间：2026-07-01 00:50

抓取时间：2026-08-02 00:26

来源机构：Microsoft Research