dataset
active
dataset:skillsbench

SkillsBench

86-task benchmark spanning 11 domains with deterministic per-task verifier, used to evaluate skill-based agent execution

Neighborhood — ranked by edge-count

Papers (1)

paper

Thinkers (1)

thinker
  • Xiangyi Li
    studies
    Author of SkillsBench, one of the three evaluation benchmarks used in this paper