Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →- 聊天区或课件区的 `interaction` 内容里包含系统按钮值，例如 `_sys_pay`、`_sys_login`、`_sys_next_chapter`。
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -12% | 0% |
interaction 内容里包含系统按钮值，例如 _sys_pay、_sys_login、_sys_next_chapter。user_input、defaultButtonText 或 readonly 把系统按钮渲染成已提交态，导致按钮变蓝、禁用或无法再次点击。markdown-flow-ui 源码时，优先在 ai-shifu 的包装层识别系统按钮内容，再覆写传给渲染器的 readonly 和默认值。user_input 回灌成默认选中态，避免按钮渲染成蓝色已选中状态。.select 或 :disabled 样式，补一层项目侧 CSS 覆写，确保按钮颜色和可点击状态符合业务预期。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 23,718 | 17,670 | -25% | 1 | 1 | 0% | 3,467 | 3,430 | -1% | 0 | 0 | — |
case-06 | pass→pass | 17,215 | 11,097 | -36% | 1 | 1 | 0% | 2,899 | 2,110 | -27% | 0 | 0 | — |
case-02 | fail→pass | 21,388 | 19,945 | -7% | 1 | 1 | 0% | 3,300 | 3,790 | +15% | 0 | 0 | — |
case-03 | fail→pass | 19,731 | 15,714 | -20% | 1 | 1 | 0% | 2,824 | 2,744 | -3% | 0 | 0 | — |
case-04 | fail→pass | 18,699 | 14,336 | -23% | 1 | 1 | 0% | 2,800 | 2,752 | -2% | 0 | 0 | — |
case-05 | pass→pass | 14,731 | 7,151 | -51% | 1 | 1 | 0% | 2,344 | 1,452 | -38% | 0 | 0 | — |
case-07 | fail→pass | 19,555 | 13,866 | -29% | 1 | 1 | 0% | 2,683 | 2,355 | -12% | 0 | 0 | — |
case-08 | fail→pass | 18,022 | 13,123 | -27% | 1 | 1 | 0% | 2,592 | 2,301 | -11% | 0 | 0 | — |
case-09 | pass→fail | 13,853 | 5,518 | -60% | 1 | 1 | 0% | 1,737 | 1,148 | -34% | 0 | 0 | — |
case-10 | pass→pass | 16,541 | 11,286 | -32% | 1 | 1 | 0% | 2,347 | 2,063 | -12% | 0 | 0 | — |
case-11 | fail→pass | 30,430 | 8,676 | -71% | 1 | 1 | 0% | 2,641 | 1,811 | -31% | 0 | 0 | — |
case-12 | fail→pass | 14,509 | 6,790 | -53% | 1 | 1 | 0% | 2,463 | 1,433 | -42% | 0 | 0 | — |
case-13 | pass→pass | 19,351 | 22,417 | +16% | 1 | 1 | 0% | 2,880 | 3,813 | +32% | 0 | 0 | — |
case-14 | fail→pass | 25,644 | 13,863 | -46% | 1 | 1 | 0% | 2,760 | 2,737 | -1% | 0 | 0 | — |
case-15 | fail→pass | 19,742 | 16,466 | -17% | 1 | 1 | 0% | 3,164 | 2,645 | -16% | 0 | 0 | — |
case-16 | fail→pass | 16,444 | 10,924 | -34% | 1 | 1 | 0% | 2,461 | 2,061 | -16% | 0 | 0 | — |
case-17 | pass→pass | 15,942 | 6,581 | -59% | 1 | 1 | 0% | 2,587 | 1,363 | -47% | 0 | 0 | — |
case-18 | pass→pass | 15,354 | 9,601 | -37% | 1 | 1 | 0% | 2,460 | 1,911 | -22% | 0 | 0 | — |
case-19 | pass→pass | 18,596 | 8,839 | -52% | 1 | 1 | 0% | 2,767 | 1,619 | -41% | 0 | 0 | — |
case-20 | fail→pass | 14,450 | 10,117 | -30% | 1 | 1 | 0% | 2,125 | 1,891 | -11% | 0 | 0 | — |
case-21 | fail→pass | 19,247 | 14,099 | -27% | 1 | 1 | 0% | 3,005 | 2,515 | -16% | 0 | 0 | — |
case-22 | fail→pass | 19,001 | 18,936 | -0% | 1 | 1 | 0% | 2,700 | 3,129 | +16% | 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 +59 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.