Chaofan Meng

SkillConsist: Detecting Inconsistencies in Agent Skills via Bidirectional Graph Alignment

This work constructs a 633-Skill benchmark from ClawHub's 500 most-downloaded public Skills and 133 Skill-Inject packages and presents SkillConsist, a model that achieves 86.85% precision, 89.03% recall, and 87.93% F1 for package-level detection, improving F1 over the best baseline by 20.43 percentage points.