AI RESEARCH
Group of Skills: Group-Structured Skill Retrieval for Agent Skill Libraries
arXiv CS.AI
•
ArXi:2605.06978v1 Announce Type: cross Skill-augmented agents increasingly rely on large reusable skill libraries, but retrieving relevant skills is not the same as presenting usable context. Existing methods typically return atomic skills or dependency-aware bundles whose internal roles remain implicit, leaving the agent to infer the execution entry point, skills, visible requirements, and failure-avoidance guidance. We