Software Engineering Skill Evaluation on SWE-Skills-Bench python-anti-patterns
4Pass RateFull Skill Text
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Full Skill TextSetting=Full Skill Text prompting2026.06 | 4 | 8.5 | 1.18 | |
| Vanilla LLMSetting=Vanilla baseline2026.06 | 3 | — | — | |
| S2LSetting=Our method2026.06 | 1 | -49.41 | -0.4 |