Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Logo 分块组装入场 + glow bloom + tagline 揭示, 适合视频片尾 / 品牌闭幕
.claude/skills/nexu-io-frame-logo-outro/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 219% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-21 | ✗→✓ | ▲ Improved | 9% | 0% |
【模板: Logo 收尾帧 (Logo Outro)】 【意图】视频结尾的品牌 reveal 帧 —— logo 分块拼装 + glow bloom + tagline 上浮 + CTA。Inspired by hyperframes logo-outro。
【画布】1920×1080, 黑色 #08090c 或品牌深色背景; 加微妙 vignette radial-gradient(...) 让中心更亮。
【布局】
filter: drop-shadow(0 0 24px <accent>40); 同时一道 shimmer mask-image 横扫 logo (500ms)。htmlanything.dev · @htmlanything · 2026, 11px uppercase letter-spacing 0.16em, 颜色 opacity 0.4, hairline 分隔。【调色 — 4 选 1, 不混用】
#08090c, accent #7c5cff (霓虹紫蓝 glow)。#0e0a08, accent #ffb547 (暖琥珀)。#0a1410, accent #5fb38a (薄荷绿)。#f1efea, accent #0a0a0b (无 neon, 走 editorial 风, glow 改成阴影)。【设计细节】
@keyframes + animation-delay; 可被 prefers-reduced-motion 关闭。Inter Tight / SF Pro Display / Manrope; 中文 Noto Sans SC weight 700。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-10 | fail→fail | 23,885 | 30,581 | +28% | 1 | 1 | 0% | 4,812 | 6,903 | +43% | 0 | 0 | — |
case-01 | fail→fail | 29,630 | 27,615 | -7% | 1 | 1 | 0% | 6,257 | 6,973 | +11% | 0 | 0 | — |
case-02 | fail→fail | 28,100 | 26,410 | -6% | 1 | 1 | 0% | 6,252 | 6,968 | +11% | 0 | 0 | — |
case-03 | fail→fail | 27,555 | 26,940 | -2% | 1 | 1 | 0% | 6,237 | 6,952 | +11% | 0 | 0 | — |
case-04 | pass→pass | 11,432 | 23,506 | +106% | 1 | 1 | 0% | 2,292 | 6,167 | +169% | 0 | 0 | — |
case-05 | pass→fail | 13,471 | 25,225 | +87% | 1 | 1 | 0% | 2,418 | 6,883 | +185% | 0 | 0 | — |
case-06 | pass→fail | 18,441 | 28,809 | +56% | 1 | 1 | 0% | 3,981 | 6,880 | +73% | 0 | 0 | — |
case-07 | fail→fail | 26,461 | 29,198 | +10% | 1 | 1 | 0% | 6,186 | 6,901 | +12% | 0 | 0 | — |
case-08 | fail→pass | 27,992 | 25,059 | -10% | 1 | 1 | 0% | 6,188 | 6,903 | +12% | 0 | 0 | — |
case-09 | fail→pass | 27,007 | 25,786 | -5% | 1 | 1 | 0% | 6,177 | 6,893 | +12% | 0 | 0 | — |
case-11 | fail→pass | 11,522 | 25,515 | +121% | 1 | 1 | 0% | 2,060 | 6,568 | +219% | 0 | 0 | — |
case-12 | fail→fail | 23,188 | 28,407 | +23% | 1 | 1 | 0% | 5,218 | 6,896 | +32% | 0 | 0 | — |
case-13 | fail→fail | 30,581 | 27,063 | -12% | 1 | 1 | 0% | 6,179 | 6,895 | +12% | 0 | 0 | — |
case-14 | fail→fail | 20,712 | 26,496 | +28% | 1 | 1 | 0% | 4,521 | 6,896 | +53% | 0 | 0 | — |
case-15 | fail→fail | 16,152 | 25,164 | +56% | 1 | 1 | 0% | 3,600 | 6,367 | +77% | 0 | 0 | — |
case-16 | pass→pass | 11,411 | 30,078 | +164% | 1 | 1 | 0% | 2,002 | 6,900 | +245% | 0 | 0 | — |
case-17 | fail→pass | 29,645 | 26,402 | -11% | 1 | 1 | 0% | 6,180 | 6,667 | +8% | 0 | 0 | — |
case-18 | fail→fail | 26,485 | 29,094 | +10% | 1 | 1 | 0% | 6,181 | 6,897 | +12% | 0 | 0 | — |
case-19 | pass→fail | 25,931 | 25,759 | -1% | 1 | 1 | 0% | 6,176 | 6,892 | +12% | 0 | 0 | — |
case-20 | pass→pass | 22,270 | 24,787 | +11% | 1 | 1 | 0% | 4,931 | 6,543 | +33% | 0 | 0 | — |
case-21 | fail→pass | 29,458 | 26,320 | -11% | 1 | 1 | 0% | 6,107 | 6,629 | +9% | 0 | 0 | — |
case-22 | fail→fail | 18,958 | 25,186 | +33% | 1 | 1 | 0% | 3,543 | 6,887 | +94% | 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 +9 percentage points is the difference between those two pass rates over the 22 comparable cases. 3 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.