Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →- 修改后台管理页的筛选区域布局、筛选项 label、查询/重置/展开按钮样式时使用。
.claude/skills/ai-shifu-admin-filter-layout/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 2% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -32% | 0% |
src/app/admin/components/AdminFilter.tsx，不要在页面内重复拼装 label、按钮组和展开/收起布局。--base-foreground、--text-sm-font-size、--font-weight-medium、--text-sm-line-height 这组 token；同一筛选区需要纵向对齐时，通过 expandedLabelClassName 或 labelClassName 传入固定宽度和右对齐，避免收起态第一行被固定 label 宽度额外顶开。Button size="sm"，左右内边距统一覆盖为 px-4；展开/收起按钮颜色使用 --base-foreground，与查询按钮间距保持 24px，文字和 chevron 间距保持 4px，hover 区域要保留可见的左内边距。AdminFilter 只负责外层布局和通用动作区。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 20,214 | 5,887 | -71% | 1 | 1 | 0% | 3,520 | 774 | -78% | 0 | 0 | — |
case-02 | fail→fail | 19,665 | 9,059 | -54% | 1 | 1 | 0% | 3,733 | 1,046 | -72% | 0 | 0 | — |
case-03 | fail→fail | 12,824 | 4,638 | -64% | 1 | 1 | 0% | 1,997 | 727 | -64% | 0 | 0 | — |
case-04 | fail→pass | 16,923 | 8,140 | -52% | 1 | 1 | 0% | 2,622 | 1,976 | -25% | 0 | 0 | — |
case-05 | fail→pass | 12,102 | 4,906 | -59% | 1 | 1 | 0% | 2,019 | 1,225 | -39% | 0 | 0 | — |
case-06 | fail→fail | 14,061 | 4,667 | -67% | 1 | 1 | 0% | 2,186 | 1,291 | -41% | 0 | 0 | — |
case-07 | fail→pass | 17,749 | 9,362 | -47% | 1 | 1 | 0% | 2,814 | 1,959 | -30% | 0 | 0 | — |
case-08 | pass→fail | 17,892 | 7,579 | -58% | 1 | 1 | 0% | 3,008 | 1,496 | -50% | 0 | 0 | — |
case-09 | pass→pass | 15,986 | 14,035 | -12% | 1 | 1 | 0% | 2,546 | 2,720 | +7% | 0 | 0 | — |
case-10 | fail→pass | 15,531 | 13,462 | -13% | 1 | 1 | 0% | 2,550 | 2,596 | +2% | 0 | 0 | — |
case-11 | fail→pass | 17,227 | 9,858 | -43% | 1 | 1 | 0% | 2,442 | 1,649 | -32% | 0 | 0 | — |
case-12 | fail→fail | 19,135 | 17,763 | -7% | 1 | 1 | 0% | 2,739 | 3,308 | +21% | 0 | 0 | — |
case-13 | pass→pass | 16,869 | 6,764 | -60% | 1 | 1 | 0% | 2,358 | 1,500 | -36% | 0 | 0 | — |
case-14 | pass→pass | 15,368 | 5,901 | -62% | 1 | 1 | 0% | 2,280 | 1,367 | -40% | 0 | 0 | — |
case-15 | fail→pass | 12,881 | 3,328 | -74% | 1 | 1 | 0% | 2,106 | 934 | -56% | 0 | 0 | — |
case-16 | fail→pass | 14,793 | 4,841 | -67% | 1 | 1 | 0% | 2,274 | 1,169 | -49% | 0 | 0 | — |
case-17 | fail→pass | 10,117 | 3,815 | -62% | 1 | 1 | 0% | 1,593 | 1,045 | -34% | 0 | 0 | — |
case-18 | fail→pass | 29,631 | 18,048 | -39% | 1 | 1 | 0% | 5,407 | 4,362 | -19% | 0 | 0 | — |
case-19 | fail→pass | 10,256 | 4,207 | -59% | 1 | 1 | 0% | 1,811 | 1,161 | -36% | 0 | 0 | — |
case-20 | pass→pass | 16,736 | 17,487 | +4% | 1 | 1 | 0% | 2,854 | 3,824 | +34% | 0 | 0 | — |
case-21 | pass→pass | 15,273 | 10,828 | -29% | 1 | 1 | 0% | 2,649 | 2,413 | -9% | 0 | 0 | — |
case-22 | pass→pass | 17,706 | 10,033 | -43% | 1 | 1 | 0% | 2,582 | 1,900 | -26% | 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, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +41 percentage points is the difference between those two pass rates over the 19 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.