Install any skill in seconds. Free to start, no credit card required.
Get Started Free →光标拖光 + 彩色像散射线 + 定向光斑, 适合视频片头逐字揭示金句
.claude/skills/nexu-io-vfx-text-cursor/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 211% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 172% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 254% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 221% | 0% |
【模板: VFX 文字光标 (Text Cursor)】 【意图】视频开场/Hero 帧 —— 光标在画布上"打字", 文字逐字浮现, 后面拖着彩色像散尾迹 + 定向光斑。Inspired by hyperframes vfx-text-cursor。
【画布】1920×1080, 背景 #06070a 暗哑黑 或 #0a0d12 (有暖偏蓝); 加微妙 vignette。
【内容】
Inter Tight / Source Sans 3 / Noto Sans SC。▍ (或细 vertical bar)。#f5f5f7, opacity 1; 即将揭示位置加 chromatic ghost: 一份 text-shadow: 2px 0 #ff3b6f, -2px 0 #00d4ff 在 reveal 瞬间, 200ms 内收敛回正常。#ff3b6f / cyan #00d4ff / amber #ffb547), 闪烁 @keyframes 1.0s 周期; 后面拖一条 60-120px 的 motion blur trail (径向渐变到透明)。【光斑 / 射线】
linear-gradient(45deg, transparent, accent20, transparent) 的细长矩形 + mix-blend-mode: screen, 不规则角度。【字段】
00:03:21 mono)。【设计细节】
Inter Tight Bold; 中文 Noto Sans SC Bold; 严禁衬线。@keyframes + JS 计时器 (setTimeout 逐字), 可被 prefers-reduced-motion 关闭 (直接显示所有字)。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 30,915 | 29,251 | -5% | 1 | 1 | 0% | 6,227 | 6,901 | +11% | 0 | 0 | — |
case-02 | fail→fail | 32,037 | 28,671 | -11% | 1 | 1 | 0% | 6,212 | 6,886 | +11% | 0 | 0 | — |
case-03 | fail→pass | 32,561 | 37,505 | +15% | 1 | 1 | 0% | 6,209 | 6,883 | +11% | 0 | 0 | — |
case-04 | fail→fail | 16,413 | 26,735 | +63% | 1 | 1 | 0% | 2,748 | 6,473 | +136% | 0 | 0 | — |
case-05 | fail→fail | 17,226 | 28,155 | +63% | 1 | 1 | 0% | 3,008 | 6,650 | +121% | 0 | 0 | — |
case-06 | pass→pass | 12,813 | 24,609 | +92% | 1 | 1 | 0% | 2,055 | 6,025 | +193% | 0 | 0 | — |
case-07 | fail→fail | 14,756 | 28,805 | +95% | 1 | 1 | 0% | 2,272 | 6,868 | +202% | 0 | 0 | — |
case-08 | fail→pass | 13,452 | 32,820 | +144% | 1 | 1 | 0% | 2,115 | 6,586 | +211% | 0 | 0 | — |
case-09 | fail→pass | 12,059 | 25,311 | +110% | 1 | 1 | 0% | 1,981 | 5,398 | +172% | 0 | 0 | — |
case-10 | fail→pass | 9,894 | 23,185 | +134% | 1 | 1 | 0% | 1,611 | 5,710 | +254% | 0 | 0 | — |
case-11 | fail→pass | 26,157 | 29,192 | +12% | 1 | 1 | 0% | 2,140 | 6,868 | +221% | 0 | 0 | — |
case-12 | fail→pass | 11,755 | 24,736 | +110% | 1 | 1 | 0% | 1,916 | 6,138 | +220% | 0 | 0 | — |
case-13 | fail→pass | 12,891 | 24,196 | +88% | 1 | 1 | 0% | 2,124 | 5,867 | +176% | 0 | 0 | — |
case-14 | fail→pass | 15,416 | 23,666 | +54% | 1 | 1 | 0% | 2,407 | 5,640 | +134% | 0 | 0 | — |
case-15 | fail→pass | 6,643 | 5,798 | -13% | 1 | 1 | 0% | 1,174 | 1,711 | +46% | 0 | 0 | — |
case-16 | fail→pass | 9,449 | 3,704 | -61% | 1 | 1 | 0% | 1,437 | 1,352 | -6% | 0 | 0 | — |
case-17 | pass→pass | 10,533 | 3,731 | -65% | 1 | 1 | 0% | 1,666 | 1,228 | -26% | 0 | 0 | — |
case-18 | fail→pass | 12,877 | 24,856 | +93% | 1 | 1 | 0% | 2,571 | 6,071 | +136% | 0 | 0 | — |
case-19 | fail→fail | 13,453 | 25,992 | +93% | 1 | 1 | 0% | 2,290 | 6,854 | +199% | 0 | 0 | — |
case-20 | pass→pass | 9,548 | 25,243 | +164% | 1 | 1 | 0% | 1,937 | 6,855 | +254% | 0 | 0 | — |
case-21 | pass→fail | 8,162 | 26,360 | +223% | 1 | 1 | 0% | 1,701 | 6,848 | +303% | 0 | 0 | — |
case-22 | pass→fail | 10,539 | 27,441 | +160% | 1 | 1 | 0% | 2,324 | 6,855 | +195% | 0 | 0 | — |
case-23 | pass→fail | 11,445 | 27,397 | +139% | 1 | 1 | 0% | 2,401 | 6,839 | +185% | 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. 23 cases were attempted. The headline lift of +35 percentage points is the difference between those two pass rates over the 23 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.