Software Engineering Skill Evaluation on SWE-Skills-Bench python-observability
7Pass CountVanilla LLM
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Vanilla LLMSetting=Vanilla baseline2026.06 | 7 | — | — | |
| Full Skill TextSetting=Full Skill Text prompting2026.06 | 3 | 18.51 | -2.16 | |
| S2LSetting=Our method2026.06 | 3 | 1.78 | -22.51 |