據 AlphaMoat 統計,截至 2026-09,UI 累計下載 4,570 次、獲 5 個星標,在全部 5,078 個 AI 技能中排名第 2,093 位,在「設計創意」分類(512 個)中排名第 204 位。
| 平台 | 累計安裝 | 累計下載 | 星標 | 評論 |
|---|---|---|---|---|
| ClawHub ↗ | 80 | 2886 | 4 | - |
| SkillHub ↗ | 106 | 4570 | 5 | — |
各平台的計數口徑不同,統計時間不一致。
主文涵蓋視覺層級、字型、色彩、對齊、元件狀態、圖示、圖像、響應式佈局、深色模式、動效和設計令牌。它提供通用建議,未必符合特定產品標準、無障礙要求或地區介面習慣;應用前應針對目標語言和本地閱讀方向調整,並依據產品規範與無障礙標準複核。
資料釋出於 2026-10-29
以下是根據技能文件生成的示例,不代表實際執行結果。
2026-09 起統計方法調整:同名技能按作者分別統計,並覆蓋全部技能。圖中虛線為調整前的月份,與之後的資料不直接比較,也不計算跨越調整的增長。
| 月份 | 累計下載 | 環比 | 星標 |
|---|---|---|---|
| 2026-09 | 4570 | — | 5 |
| 2026-08 | 3415 | +27.9% | 5 |
| 2026-07 | 2671 | +24.4% | 5 |
| 2026-06 | 2147 | +14.9% | 4 |
| 2026-05 | 1869 | — | 4 |
| 2026-03 | 1193 | — | 4 |
以下是作者釋出的原始檔案,按純文字顯示,本站未做修改或稽核;其中的連結不可點選,命令請核實後再執行。抓取於 2026-10-23。
---
name: UI
description: Design clear, consistent, and visually polished user interfaces.
metadata: {"clawdbot":{"emoji":"🎨","os":["linux","darwin","win32"]}}
---
## Visual Hierarchy
- One focal point per screen—eye knows where to go first
- Size, color, weight establish importance—primary action most prominent- Group related elements—proximity implies relationship - White space is not wasted space—breathing room aids scanning ## Typography - Maximum 2-3 font families—more creates visual noise - Clear size scale: title > heading > body > caption—distinct steps, not gradual - Line height 1.4-1.6 for body text—too tight or loose hurts readability - Line length 45-75 characters—prevents eye fatigue - Left-align body text—centered only for short headings ## Color Usage - Primary color for primary actions—one dominant brand color - Semantic colors consistent: red=error, green=success, yellow=warning - Don't rely on color alone—add icons, text, patterns for accessibility - Neutral palette for most UI—color for emphasis, not everywhere - Test color blindness scenarios—8% of men affected ## Spacing System - Use consistent scale: 4px, 8px, 16px, 24px, 32px, 48px - Apply same spacing for same relationships—all card padding equal - More space around groups than within—visual grouping through proximity - Generous padding on touch targets—44px minimum for mobile ## Alignment - Grid system for consistency—8px or 4px base grid - Align to invisible lines—elements share edges, not scattered - Left edge strongest for LTR—anchor content predictably - Optical alignment when needed—visual center differs from mathematical ## Component States - Default, hover, active, focus, disabled—all states designed - Focus state visible and clear—keyboard users need this - Disabled looks disabled—reduced opacity, no pointer cursor - Loading state replaces content—not just overlay - Error state in context—red border, inline message ## Icons - Consistent style throughout—don't mix outlined and filled - Recognizable at small sizes—simple shapes work better - Labels when meaning ambiguous—icon + text clearer than icon alone - Touch target larger than visual icon—44px tap area, 24px icon ## Imagery - Consistent aspect ratios—don't stretch or skew - Fallback for failed loads—placeholder, not broken image - Alt text for content images—decorative images alt="" - Compress appropriately—quality vs file size balance ## Responsive Design - Design for smallest screen first—enhance for larger - Breakpoints based on content—not arbitrary device widths - Touch targets larger on touch screens—hover states only on desktop - Consider landscape orientation—especially for tablets ## Dark Mode - Not just color inversion—redesign depth and emphasis - Reduce contrast slightly—pure white on black strains eyes - Shadows don't work same—use lighter surfaces for elevation - Test all states—errors, success, charts, images - Respect system preference—but allow override ## Motion and Animation - Duration 150-300ms for transitions—fast but perceptible - Ease-out for entering—starts fast, settles in - Ease-in for exiting—accelerates out of view - Consistent timing across similar interactions - Purpose: guide attention, show relationships, provide feedback ## Design Tokens - Define tokens for colors, spacing, typography—single source of truth - Semantic naming: `color-error` not `color-red` - Enables theming and dark mode—swap token values - Scales with product—change once, update everywhere ## Common Mistakes - Too many font sizes—stick to the scale - Inconsistent spacing—creates unpolished feel - Low contrast text—4.5:1 minimum for accessibility - Buttons that don't look clickable—affordance matters - Different styles for same component—cards should match cards
ContextWeave 架構圖一鍵生成
用 Nano Banana Pro 生成或編輯圖片的指令碼技能。
用 whisper 命令列工具在本機將音訊轉成文字或進行翻譯。
透過 MCP 服務生成與編輯圖表的技能。
面向介面設計的思路與落地建議技能,覆蓋佈局、互動流程、設計令牌與前端改動。
面向多家影片模型 API 的 AI 影片生成與編排技能。
用 ffmpeg 從影片裡取出畫面幀,或生成便於快速檢視的縮圖。
用本地 Whisper 把音訊檔案轉成文字,模型下載完成後可完全離線執行。
將使用者明確糾正與自我檢視整理成分層記憶,並依使用情況升級或封存。
在保留公式、格式和活頁簿結構的前提下建立、檢查或編輯 Excel 檔案。
處理 Word 與 DOCX 文件的作業指南,著重維持樣式、編號與版面。
把資料庫、表格與匯出的原始資料變成可決策結論的資料分析技能。
協助編輯 PowerPoint 簡報,並在交付前檢查版面是否正常呈現。
資料月份:2026-09 · 下載/星標/安裝數聚合自公開技能註冊庫(ClawHub、SkillHub),按月更新。口徑說明見資料方法。