Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →日本語翻訳：このファイルは nanoclaw-repl 用の日本語翻訳が必要です
.claude/skills/affaan-m-nanoclaw-repl/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -37% | 0% |
在运行或扩展 scripts/claw.js 时使用此技能。
/model 进行模型切换/load 进行动态技能加载/branch 进行会话分支/search 进行跨会话搜索/compact 进行历史压缩/export 导出为 md/json/txt 格式/metrics 查看会话指标| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-16 | fail→pass | 11,510 | 4,434 | -61% | 1 | 1 | 0% | 1,652 | 926 | -44% | 0 | 0 | — |
case-01 | fail→pass | 19,515 | 16,443 | -16% | 1 | 1 | 0% | 3,287 | 3,100 | -6% | 0 | 0 | — |
case-02 | fail→pass | 14,294 | 8,893 | -38% | 1 | 1 | 0% | 2,302 | 1,678 | -27% | 0 | 0 | — |
case-03 | fail→pass | 8,481 | 5,507 | -35% | 1 | 1 | 0% | 1,385 | 1,288 | -7% | 0 | 0 | — |
case-04 | pass→pass | 5,753 | 1,729 | -70% | 1 | 1 | 0% | 898 | 430 | -52% | 0 | 0 | — |
case-05 | pass→pass | 8,431 | 3,133 | -63% | 1 | 1 | 0% | 1,395 | 492 | -65% | 0 | 0 | — |
case-06 | fail→pass | 5,244 | 2,145 | -59% | 1 | 1 | 0% | 779 | 489 | -37% | 0 | 0 | — |
case-07 | fail→pass | 11,062 | 2,019 | -82% | 1 | 1 | 0% | 1,828 | 508 | -72% | 0 | 0 | — |
case-08 | pass→pass | 10,232 | 3,308 | -68% | 1 | 1 | 0% | 1,600 | 676 | -58% | 0 | 0 | — |
case-09 | pass→pass | 9,021 | 2,273 | -75% | 1 | 1 | 0% | 1,397 | 540 | -61% | 0 | 0 | — |
case-10 | fail→pass | 5,426 | 2,019 | -63% | 1 | 1 | 0% | 893 | 490 | -45% | 0 | 0 | — |
case-11 | pass→pass | 7,890 | 3,320 | -58% | 1 | 1 | 0% | 1,009 | 639 | -37% | 0 | 0 | — |
case-12 | fail→pass | 7,536 | 2,877 | -62% | 1 | 1 | 0% | 1,058 | 556 | -47% | 0 | 0 | — |
case-13 | pass→pass | 6,240 | 2,961 | -53% | 1 | 1 | 0% | 902 | 671 | -26% | 0 | 0 | — |
case-14 | fail→pass | 6,995 | 3,255 | -53% | 1 | 1 | 0% | 1,000 | 667 | -33% | 0 | 0 | — |
case-15 | pass→pass | 11,709 | 5,416 | -54% | 1 | 1 | 0% | 1,763 | 1,151 | -35% | 0 | 0 | — |
case-17 | fail→pass | 11,255 | 4,148 | -63% | 1 | 1 | 0% | 1,690 | 836 | -51% | 0 | 0 | — |
case-18 | fail→pass | 19,832 | 3,609 | -82% | 1 | 1 | 0% | 1,761 | 376 | -79% | 0 | 0 | — |
case-19 | pass→pass | 9,337 | 2,967 | -68% | 1 | 1 | 0% | 1,529 | 660 | -57% | 0 | 0 | — |
case-20 | pass→pass | 3,280 | 2,294 | -30% | 1 | 1 | 0% | 474 | 508 | +7% | 0 | 0 | — |
case-21 | pass→pass | 25,196 | 9,815 | -61% | 1 | 1 | 0% | 2,141 | 2,023 | -6% | 0 | 0 | — |
case-22 | pass→pass | 7,542 | 5,365 | -29% | 1 | 1 | 0% | 1,390 | 1,060 | -24% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +50 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.