1 paper · 1 filter
Tong Bai, Zhenglin Wan, Pengfei Zhou +3
As LLM agents adopt large skill libraries, selecting the right subset becomes a structural problem rather than a similarity-matching one: skills depend on, conflict with, specializ…