AI RESEARCH
SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks
arXiv CS.AI
•
ArXi:2602.12670v2 Announce Type: replace Agent Skills are structured packages of procedural knowledge that augment LLM agents at inference time. Despite rapid adoption, there is no standard way to measure whether they actually help. We present SkillsBench, a benchmark of 86 tasks across 11 domains paired with curated Skills and deterministic verifiers. We test 7 agent-model configurations over 7,308 trajectories.