From Raw Experience to Skill Consumption: A Systematic Study of Model-Generated Agent Skills
The study explores the lifecycle of model-generated agent skills, focusing on experience generation, skill extraction, and skill consumption. It reveals that while model-generated skills are generally beneficial, they can also lead to negative transfer. The authors propose a meta-skill to enhance skill extraction and improve overall skill quality across various domains.
- ▪Model-generated skills are structured procedural artifacts that improve agent performance by reusing past experiences.
- ▪The study identifies that skill utility is not dependent on model scale or baseline task strength.
- ▪A meta-skill is introduced to guide skill extraction towards features that enhance utility and reduce negative transfer.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.23899 (cs) [Submitted on 22 May 2026] Title:From Raw Experience to Skill Consumption: A Systematic Study of Model-Generated Agent Skills Authors:Zisu Huang, Jingwen Xu, Yifan Yang, Ziyang Gong, Qihao Yang, Muzhao Tian, Xiaohua Wang, Changze Lv, Xuemei Gao, Qi Dai, Bei Liu, Kai Qiu, Xue Yang, Dongdong Chen, Xiaoqing Zheng, Chong Luo View a PDF of the paper titled From Raw Experience to Skill Consumption: A Systematic Study of Model-Generated Agent Skills, by Zisu Huang and 15 other authors View PDF HTML (experimental) Abstract:Language agents increasingly improve by reusing \emph{skills} -- structured procedural artifacts distilled from past experience.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
