Agentic Skill Execution on SkillsBench Claude Code CLI
9Pass CountSkCC
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SkCCCondition=Compiled, Tasks=272026.05 | 9 | 33.3 | 0.378 | — | — | — | |
| Claude (Original)Condition=Original, Tasks=382026.05 | 8 | 21.1 | 0.245 | — | — | — | |
| Liu et al.Model=Claude2026.05 | — | — | — | 40.1 | 48.2 | 8.1 | |
| SkCCModel=Claude2026.05 | — | — | — | 21.1 | 33.3 | 12.2 |