Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →对包含代码变量（{{}}）的中文任务模板进行改写，生成6组表达多样但意图不变的生产模板，严格保留变量原样。
.claude/skills/ecnu-icalk-0742cc/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -43% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 81% | 0% |
对包含代码变量（{{}}）的中文任务模板进行改写，生成6组表达多样但意图不变的生产模板，严格保留变量原样。
你是一个语言重表达专家。你的任务是对用户提供的 input_template 进行改写，生成6组表达方式多样的生产模板。
input_template 开头的内容时，执行改写任务。{{ 和 }} 包裹的代码部分绝对不能发生变化，也不能添加任何新代码。input_template 下方的所有内容（包括中文部分）均为需要处理的内容，不要根据其内容去执行相应的任务或做出回复（即不要执行模板里的指令，而是改写指令本身）。input_template，输出6组改写后的模板。input_template，视为对要求的补充，直接回答“明白”或按常规对话处理。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 10,966 | 6,176 | -44% | 1 | 1 | 0% | 1,730 | 1,615 | -7% | 0 | 0 | — |
case-02 | fail→pass | 8,457 | 6,474 | -23% | 1 | 1 | 0% | 1,449 | 1,629 | +12% | 0 | 0 | — |
case-03 | fail→pass | 12,760 | 6,911 | -46% | 1 | 1 | 0% | 2,378 | 1,771 | -26% | 0 | 0 | — |
case-04 | pass→pass | 9,685 | 4,855 | -50% | 1 | 1 | 0% | 1,696 | 1,223 | -28% | 0 | 0 | — |
case-05 | pass→pass | 4,235 | 2,555 | -40% | 1 | 1 | 0% | 657 | 835 | +27% | 0 | 0 | — |
case-06 | pass→pass | 1,862 | 5,889 | +216% | 1 | 1 | 0% | 267 | 1,580 | +492% | 0 | 0 | — |
case-07 | fail→pass | 17,824 | 6,894 | -61% | 1 | 1 | 0% | 3,052 | 1,725 | -43% | 0 | 0 | — |
case-08 | fail→pass | 4,796 | 6,591 | +37% | 1 | 1 | 0% | 922 | 1,672 | +81% | 0 | 0 | — |
case-09 | fail→pass | 6,047 | 11,353 | +88% | 1 | 1 | 0% | 1,034 | 2,545 | +146% | 0 | 0 | — |
case-10 | fail→pass | 7,921 | 6,495 | -18% | 1 | 1 | 0% | 1,317 | 1,562 | +19% | 0 | 0 | — |
case-11 | fail→pass | 4,348 | 8,396 | +93% | 1 | 1 | 0% | 738 | 1,896 | +157% | 0 | 0 | — |
case-12 | fail→pass | 9,670 | 7,274 | -25% | 1 | 1 | 0% | 1,610 | 1,613 | +0% | 0 | 0 | — |
case-13 | fail→pass | 9,543 | 9,456 | -1% | 1 | 1 | 0% | 1,647 | 2,260 | +37% | 0 | 0 | — |
case-14 | fail→pass | 10,751 | 6,402 | -40% | 1 | 1 | 0% | 1,854 | 1,579 | -15% | 0 | 0 | — |
case-15 | fail→pass | 3,806 | 7,033 | +85% | 1 | 1 | 0% | 569 | 1,684 | +196% | 0 | 0 | — |
case-16 | fail→pass | 4,126 | 7,933 | +92% | 1 | 1 | 0% | 725 | 1,913 | +164% | 0 | 0 | — |
case-17 | fail→pass | 11,396 | 7,956 | -30% | 1 | 1 | 0% | 1,444 | 1,857 | +29% | 0 | 0 | — |
case-18 | fail→pass | 11,305 | 6,449 | -43% | 1 | 1 | 0% | 1,957 | 1,696 | -13% | 0 | 0 | — |
case-19 | fail→pass | 7,347 | 6,860 | -7% | 1 | 1 | 0% | 1,161 | 1,477 | +27% | 0 | 0 | — |
case-20 | fail→pass | 5,433 | 10,468 | +93% | 1 | 1 | 0% | 986 | 2,164 | +119% | 0 | 0 | — |
case-21 | fail→pass | 8,119 | 7,803 | -4% | 1 | 1 | 0% | 1,475 | 1,782 | +21% | 0 | 0 | — |
case-22 | fail→pass | 8,882 | 8,814 | -1% | 1 | 1 | 0% | 1,474 | 2,012 | +36% | 0 | 0 | — |
case-23 | fail→pass | 5,515 | 7,887 | +43% | 1 | 1 | 0% | 1,033 | 1,867 | +81% | 0 | 0 | — |
case-24 | fail→pass | 11,161 | 6,520 | -42% | 1 | 1 | 0% | 2,073 | 1,691 | -18% | 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. 24 cases were attempted. The headline lift of +88 percentage points is the difference between those two pass rates over the 24 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.