AGENT SKILL

social-youtube-thumbnail

High-contrast 1280×720 YouTube video thumbnail that stays legible at small sizes, keeps the timestamp corner clear, and exports under YouTube's 2 MB limit.

License 许可
MIT
最近更新
2026-11-14
Stars
★ 1.1k
Forks
104

01 前言 Front Matter

od
{ "mode": "prototype", "craft": { "requires": [ "typography", "color", "anti-ai-slop" ] }, "preview": { "type": "html", "entry": "index.html", "reload": "debounce-100" }, "surface": "web", "platform": "desktop", "scenario": "marketing", "design_system": { "requires": false }, "example_prompt": "Use the YouTube Thumbnail template to design a thumbnail for my video. Make the promise of the video readable in under a second: one short hook line, one strong focal subject, high contrast. Use real content from my brief and avoid lorem ipsum or placeholder images." }
name
social-youtube-thumbnail
tags
[ "youtube", "thumbnail", "social", "video", "cover", "封面" ]
emoji
▶️
en_name
YouTube Thumbnail
zh_name
YouTube 缩略图
category
card
featured
45
scenario
marketing
aspect_hint
1280×720 (16:9)
description
High-contrast 1280×720 YouTube video thumbnail that stays legible at small sizes, keeps the timestamp corner clear, and exports under YouTube's 2 MB limit.
en_description
High-contrast 1280×720 YouTube video thumbnail that stays legible at small sizes, keeps the timestamp corner clear, and exports under YouTube's 2 MB limit.
zh_description
1280×720 高对比 YouTube 缩略图, 小尺寸可读, 避开右下角时长标记, 导出小于 2 MB

03 SKILL.md 正文

YouTube Thumbnail

Intent. A single, fixed-size 1280×720 frame that sells one video. It is judged at postage-stamp size in a crowded feed, so it must communicate its hook in under a second.

Canvas

  • The root element is exactly width: 1280px; height: 720px; overflow: hidden;. Nothing may extend past it. Put data-od-card on the root so the export crops to it exactly.
  • Single self-contained HTML file. Inline SVG for icons and shapes. Web fonts only from Google Fonts. No external image URLs unless the user supplied the image. Otherwise build the subject from inline SVG, CSS gradients and shapes, or typography. Don't use emoji as the subject or any key visual. Emoji come from the exporting machine's fonts and show as empty boxes where none is installed (common on Linux and CI).

Composition

  • One focal subject occupying 35–55% of the frame (a face or reaction, the product, a big number, or an object drawn in inline SVG), placed on a third line, never dead center with text on top of it.
  • One hook line, 2–5 words, set at 96–160px display weight (800–900) with tight leading (0.95–1.05). It must add to the video title, not repeat it. At most 2 lines.
  • An optional small secondary tag (a pill or badge, 28–40px): "NEW", "v2", "vs", "3 MIN", a price, a stat.
  • High contrast: a text/background contrast ratio of at least 7:1 on the hook. Use a 6–12px stroke or a hard drop shadow (text-shadow: 0 6px 0 rgba(0,0,0,.35)) when text sits over imagery.
  • 2–3 saturated colors at most, plus black or white. One accent color, used once or twice.

Safe areas (hard rules)

  • Bottom-right 240×90px is reserved: YouTube overlays the video duration there. No text, faces, or key detail in it.
  • Keep all text at least 48px from every edge.
  • Don't put anything important in the bottom 12% across the full width either. Hover progress bars and "watch later" overlays cover it.

Legibility check

Before finishing, mentally scale the frame to 168×94px (the smallest feed size). The hook must still be readable, and the subject still recognizable. If the hook is over 5 words or under ~90px, cut words, not size.

Brand and design system

If an active design system is present, take its accent color and display typeface, but keep the thumbnail's contrast and size rules above. Thumbnails always outrank brand subtlety.

Export

The file must upload under 2 MB. After registering, export with export_open_design_artifact using width: 1280, height: 720, selector: "[data-od-card]", and maxBytes: 2000000. The tool switches to JPEG automatically if a PNG would be too large.

04 同仓库更多 Skills

feimacode
feimacode/open-design-agent-kit

youtube-clipper

YouTube clip generation and editing with automated workflows — pull source video, slice highlights, add captions, and export.

★ 1.1k
feimacode
feimacode/open-design-agent-kit

writing-guidelines

Review docs/prose for Writing Guidelines compliance. Use when asked to "review my docs", "check writing style", "audit prose", "review docs voice and tone", or "check this page against the writing handbook".

★ 1.1k
feimacode
feimacode/open-design-agent-kit

wpds

WordPress Design System. Apply WordPress's official design tokens, typography, and component patterns to themes and sites.

★ 1.1k
feimacode
feimacode/open-design-agent-kit

weread-year-in-review-video-template

WeRead-inspired HyperFrames video template for vertical annual reading reports, personal reading dashboards, book-note recaps, and shareable year-in-review stories. Use when users want a 9:16 HTML-to-MP4 reading report with warm paper texture, editorial Chinese typography, book-page metaphors, data highlights, and deterministic motion.

★ 1.1k
feimacode
feimacode/open-design-agent-kit

web-design-guidelines

Review UI code for Web Interface Guidelines compliance by the Vercel engineering team. Covers layout, typography, color, motion, and accessibility for product UI.

★ 1.1k
feimacode
feimacode/open-design-agent-kit

web-clone

网站复刻 / 克隆方法论。USE WHEN 用户说 复刻网站、克隆网站、clone website、抄个站、仿站、 照着这个站做一个、reproduce site、还原某个网页效果、把这个站搬下来改成我的、 复刻某个交互/WebGL/Canvas/Three.js 效果。提供「先拿真源码 → 判路径 → 逆向拆解 → 搭工程 → 替换内容」的可移植决策树,覆盖静态站 / React-Vue-Next 内容站 / WebGL-Canvas 重前端站三大分支,并强制核对任何 AI 二手分析里的可执行代码。

★ 1.1k