Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Claude Code 老師建專案工作模式。說「初始化班級工具」「老師建專案」時載入。
.claude/skills/mathruffian-dot-cc-workspace/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -15% | 0% |
07-workflow-skills開工：讀 CLAUDE.md → 讀 Obsidian 工作筆記 → git status → 回報下一步 收工：檢查敏感資料 → 更新筆記 → git commit/push → chezmoi 同步
回報：已建立的檔案清單、GitHub repo URL、Obsidian 筆記路徑。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 8,811 | 5,058 | -43% | 1 | 1 | 0% | 1,541 | 828 | -46% | 0 | 0 | — |
case-02 | fail→fail | 10,814 | 4,139 | -62% | 1 | 1 | 0% | 1,711 | 790 | -54% | 0 | 0 | — |
case-03 | fail→fail | 11,519 | 12,055 | +5% | 1 | 1 | 0% | 1,941 | 2,221 | +14% | 0 | 0 | — |
case-04 | fail→pass | 10,152 | 5,692 | -44% | 1 | 1 | 0% | 1,714 | 1,088 | -37% | 0 | 0 | — |
case-05 | fail→pass | 16,002 | 20,684 | +29% | 1 | 1 | 0% | 3,194 | 2,541 | -20% | 0 | 0 | — |
case-11 | fail→fail | 16,534 | 3,021 | -82% | 1 | 1 | 0% | 2,537 | 556 | -78% | 0 | 0 | — |
case-06 | fail→pass | 17,124 | 11,282 | -34% | 1 | 1 | 0% | 2,488 | 2,085 | -16% | 0 | 0 | — |
case-07 | fail→pass | 17,928 | 8,643 | -52% | 1 | 1 | 0% | 2,570 | 1,554 | -40% | 0 | 0 | — |
case-08 | fail→fail | 11,613 | 4,900 | -58% | 1 | 1 | 0% | 1,914 | 910 | -52% | 0 | 0 | — |
case-09 | pass→pass | 10,152 | 4,981 | -51% | 1 | 1 | 0% | 1,568 | 987 | -37% | 0 | 0 | — |
case-10 | fail→pass | 11,142 | 7,938 | -29% | 1 | 1 | 0% | 1,735 | 1,483 | -15% | 0 | 0 | — |
case-12 | fail→fail | 8,912 | 2,169 | -76% | 1 | 1 | 0% | 1,287 | 526 | -59% | 0 | 0 | — |
case-13 | pass→pass | 14,266 | 9,507 | -33% | 1 | 1 | 0% | 2,315 | 1,753 | -24% | 0 | 0 | — |
case-14 | fail→pass | 11,306 | 2,743 | -76% | 1 | 1 | 0% | 1,859 | 535 | -71% | 0 | 0 | — |
case-15 | fail→fail | 10,502 | 2,289 | -78% | 1 | 1 | 0% | 1,704 | 504 | -70% | 0 | 0 | — |
case-16 | pass→pass | 11,698 | 4,577 | -61% | 1 | 1 | 0% | 1,776 | 808 | -55% | 0 | 0 | — |
case-17 | fail→fail | 14,780 | 4,117 | -72% | 1 | 1 | 0% | 2,556 | 849 | -67% | 0 | 0 | — |
case-18 | fail→pass | 14,479 | 9,227 | -36% | 1 | 1 | 0% | 2,210 | 1,670 | -24% | 0 | 0 | — |
case-19 | pass→pass | 16,329 | 5,647 | -65% | 1 | 1 | 0% | 2,420 | 1,068 | -56% | 0 | 0 | — |
case-20 | pass→fail | 18,128 | 12,223 | -33% | 1 | 1 | 0% | 3,132 | 2,338 | -25% | 0 | 0 | — |
case-21 | pass→pass | 15,908 | 15,701 | -1% | 1 | 1 | 0% | 2,956 | 3,065 | +4% | 0 | 0 | — |
case-22 | pass→pass | 16,242 | 14,113 | -13% | 1 | 1 | 0% | 2,983 | 2,993 | +0% | 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 +27 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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.