Software Engineering Skill Evaluation on SWE-Skills-Bench (tdd-workflow)
4Pass RateS2L
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| S2LSetting=Our method2026.06 | 4 | 5.05 | 5.94 | |
| Vanilla LLMSetting=Vanilla baseline2026.06 | 1 | — | — | |
| Full Skill TextSetting=Full Skill Text prompting2026.06 | 1 | 11.11 | 0 |