Downstream task execution on AgentSkillOS
84.59ScoreSKILLRAE
Evaluation Results
| Method | Links | |
|---|---|---|
| SKILLRAE2026.05 | 84.59 | |
| AgentSkillOS2026.05 | 83.5 | |
| SkillRouter2026.05 | 82.3 | |
| LLM-based Retrieval2026.05 | 82.06 | |
| Benchmark-native Curated Skills2026.05 | 78.45 | |
| Vanilla Retrieval2026.05 | 74.85 |