Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →16 列网格 + 单一饱和 accent + 22 个锁死版面 (Klein Blue / Lemon / Mint / Safety Orange)
.claude/skills/nexu-io-deck-swiss-international/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 21% | 0% |
【模板: 瑞士国际主义 Deck (Swiss International)】 【意图】事实、产品、分析、方法论表达。极度冷静、理性、学院派, 没有任何手绘 / 噪点 / 装饰。Inspired by op7418/guizang-ppt-skill Style B。
【主题】只能从下面 4 套二选一, 不许混用、不许改 hex:
#002FA7, paper #fafaf8, ink #0a0a0a. 商业 / AI / 设计场景。#FFD500, paper #f7f5ee (淡奶油), ink #0a0a0a. 年轻 / 零售 / 体育。文字必须用黑色 (不能白色)。#C5E803, paper #f7f5ee, ink #0a0a0a. 可持续 / 科技初创 / Gen-Z 品牌。文字必须用黑色。#FF6B35, paper #f7f5ee, ink #0a0a0a. 工业 / 汽车 / 紧急消息。文字用白色 + bold ≥ 600。【布局 — 22 个可复用版式池, 不许新增或改造版式; 数量由内容决定, 把【用户内容】完整覆盖完为止 (短内容 6-10 张起步, 长内容应远超此范围, 同一版式可在不同章节重复使用)】
【设计细节 — 绝对铁律】
border-radius: 0。圆角 = 立刻违反。grid-template-columns: repeat(16, 1fr); gap: 0。№N/N 右下, topic 标签左下。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 28,923 | 26,446 | -9% | 1 | 1 | 0% | 6,213 | 7,537 | +21% | 0 | 0 | — |
case-02 | fail→fail | 29,139 | 27,155 | -7% | 1 | 1 | 0% | 6,211 | 7,535 | +21% | 0 | 0 | — |
case-03 | fail→pass | 28,490 | 27,068 | -5% | 1 | 1 | 0% | 6,202 | 7,526 | +21% | 0 | 0 | — |
case-04 | fail→pass | 29,166 | 28,416 | -3% | 1 | 1 | 0% | 6,170 | 7,495 | +21% | 0 | 0 | — |
case-05 | fail→pass | 27,247 | 27,081 | -1% | 1 | 1 | 0% | 6,170 | 7,494 | +21% | 0 | 0 | — |
case-06 | fail→pass | 27,103 | 26,720 | -1% | 1 | 1 | 0% | 6,171 | 7,496 | +21% | 0 | 0 | — |
case-07 | pass→pass | 28,008 | 27,786 | -1% | 1 | 1 | 0% | 6,171 | 7,495 | +21% | 0 | 0 | — |
case-08 | fail→pass | 28,172 | 30,442 | +8% | 1 | 1 | 0% | 6,171 | 7,495 | +21% | 0 | 0 | — |
case-09 | fail→pass | 26,933 | 25,157 | -7% | 1 | 1 | 0% | 6,165 | 7,489 | +21% | 0 | 0 | — |
case-10 | fail→fail | 26,482 | 27,859 | +5% | 1 | 1 | 0% | 6,172 | 7,496 | +21% | 0 | 0 | — |
case-11 | fail→fail | 27,954 | 27,075 | -3% | 1 | 1 | 0% | 6,173 | 7,497 | +21% | 0 | 0 | — |
case-12 | fail→fail | 31,797 | 28,624 | -10% | 1 | 1 | 0% | 6,167 | 7,491 | +21% | 0 | 0 | — |
case-13 | fail→pass | 29,698 | 24,269 | -18% | 1 | 1 | 0% | 6,173 | 7,496 | +21% | 0 | 0 | — |
case-14 | fail→fail | 27,302 | 27,663 | +1% | 1 | 1 | 0% | 6,176 | 7,500 | +21% | 0 | 0 | — |
case-15 | fail→fail | 29,334 | 27,517 | -6% | 1 | 1 | 0% | 6,164 | 7,488 | +21% | 0 | 0 | — |
case-16 | fail→fail | 29,141 | 28,674 | -2% | 1 | 1 | 0% | 6,167 | 7,491 | +21% | 0 | 0 | — |
case-17 | pass→fail | 26,087 | 27,209 | +4% | 1 | 1 | 0% | 5,971 | 7,490 | +25% | 0 | 0 | — |
case-18 | pass→fail | 28,707 | 28,187 | -2% | 1 | 1 | 0% | 6,169 | 7,493 | +21% | 0 | 0 | — |
case-19 | pass→pass | 28,572 | 27,854 | -3% | 1 | 1 | 0% | 6,170 | 7,495 | +21% | 0 | 0 | — |
case-20 | pass→fail | 27,239 | 27,940 | +3% | 1 | 1 | 0% | 6,174 | 7,498 | +21% | 0 | 0 | — |
case-21 | pass→fail | 28,829 | 26,176 | -9% | 1 | 1 | 0% | 6,172 | 7,496 | +21% | 0 | 0 | — |
case-22 | pass→pass | 27,535 | 27,559 | +0% | 1 | 1 | 0% | 6,174 | 7,498 | +21% | 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 +18 percentage points is the difference between those two pass rates over the 22 comparable cases. 4 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.