SkillX: Automatically Constructing Skill Knowledge Bases for Agents

Source: arXiv Authors: Chenxi Wang, Zhuoyun Yu, Xin Xie, Wuguannan Yao, Runnan Fang, Shuofei Qiao, Kexin Cao, Guozhou Zheng, Xiang Qi, Peng Zhang, Shumin Deng Date: 2026-04-06 Primary category: cs.CL All categories: cs.CL, cs.AI, cs.IR, cs.LG, cs.MA

Abstract

SkillX builds a plug-and-play skill knowledge base from trajectories using a hierarchy of strategic plans, functional skills, and atomic skills. It iteratively refines skills with execution feedback and proactively expands the library with newly validated capabilities. The paper is useful for harnesses that want workflow evolution to produce reusable procedural assets rather than one-off improvements.