← 返回 Skills 市场
openlark

UI Scanner — Website Design System Extraction

作者 OpenLark · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
64
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install ui-scanner
功能描述
Given a website URL, crawl and analyze its visual design system — identify design style, color system, typography, component styles, and UI patterns. Outputs...
使用说明 (SKILL.md)

UI Scanner — Website Design System Extraction

Given a website URL, automatically crawl the page and reverse-engineer its visual design system, outputting a structured design specification document.

Use Cases

Suitable for competitive design analysis, UI replication, design system reverse engineering, and style transfer.

Usage

Provide a target website URL; the skill performs crawl → analyze → structured output. Result written to {domain}_design.md.

Workflow

URL input → Page crawl → Visual element classification → Style extraction → Structured output → Write file

Output Format

Output file uses standard YAML frontmatter + Markdown structure:

Design Overview

---
version: alpha
name: {brand name}
description: {1-3 sentence brand visual style description, including base color, core tone, font scheme}
colors:
  primary: "{primary brand color HEX}"
  primary-active: "{primary hover/active HEX}"
  ink: "{darkest text color HEX}"
  body: "{body text color HEX}"
  body-strong: "{emphasis text color HEX}"
  muted: "{secondary text color HEX}"
  muted-soft: "{weakest text color HEX}"
  hairline: "{border color HEX}"
  hairline-soft: "{weak border color HEX}"
  hairline-strong: "{emphasis border color HEX}"
  canvas: "{page background HEX}"
  canvas-soft: "{weak background HEX}"
  surface-card: "{card background HEX}"
  surface-strong: "{emphasis surface HEX}"
  on-primary: "{text color on primary HEX}"
  semantic-error: "{error color HEX}"
  semantic-success: "{success color HEX}"
typography:
  display-mega:
    fontFamily: "{font stack}"
    fontSize: {px}
    fontWeight: {numeric}
    lineHeight: {numeric}
    letterSpacing: {px}
  display-lg: {same}
  display-md: {same}
  body-md: {same}
  body-sm: {same}
  caption: {same}
  code: {same}
  button: {same}
  nav-link: {same}
rounded:
  none: 0px
  sm: {px}
  md: {px}
  lg: {px}
  xl: {px}
  pill: {px}
spacing:
  xs: {px}
  sm: {px}
  base: {px}
  lg: {px}
  xl: {px}
  xxl: {px}
  section: {px}
components:
  top-nav: {common page component style map}
  button-primary: {backgroundColor, textColor, typography, rounded, padding, height}
  button-secondary: {same format}
  hero-band: {same}
  feature-card: {same}
  code-block: {if applicable}
  text-input: {if applicable}
  footer: {same}
---

Analysis Dimensions

1. Design Style Description

After crawling, assess from these dimensions:

  • Brand tone: Warm/cool/neutral, pure white/cream/dark base
  • Visual language: Skeuomorphic/flat/neumorphic/minimalist/editorial/magazine/tech
  • Brand color strategy: Single/dual/multi-color system, restrained vs expansive usage
  • Depth rendering: Shadow levels, borders, layering
  • Typography tone: Weight strategy (bold brand / 400 magazine), letter spacing (tight/loose)
  • Code/technical rendering: Code block style, monospace font strategy
  • UI texture: Border radius strategy, edge treatment, card style

Output a 3-5 sentence description in the description field of the frontmatter.

2. Color System

Extract the following color categories:

Category Description
Primary brand Main CTA and brand identity color
Primary hover Primary hover/active variant
Text scale 4-5 levels from darkest to lightest
Border scale 2-3 levels of hairline/border colors
Page background Canvas, card surface, emphasis surface hierarchy
Semantic colors Error red / success green (if visible on page)
Special colors Timeline, tags, auxiliary colors for special components

Extract precise HEX values from CSS (inline/computed/CSS variables/design system).

3. Typography

Extract the following typography levels:

Level Location
Display mega H1 / Hero heading
Display lg/md/sm Large/medium/small display text
Title md/sm Heading text
Body md/sm Body text
Caption Auxiliary text
Code Code areas
Button Button text
Nav link Navigation links

Record: fontFamily, fontSize, fontWeight, lineHeight, letterSpacing.

4. Spacing & Border Radius

Dimension Extraction Method
Border radius Buttons, cards, inputs, badges border-radius
Spacing Paragraph gaps, grid gaps, section spacing
Page rhythm Content area padding, section divider spacing

5. Component Styles

Extract styles for common visible components:

Component Extracted Properties
Top nav Height, background, text color, link style
Button primary Background, text, radius, padding, height
Button secondary Same as above
Feature card Background, radius, padding
Pricing card (If present) background, radius, padding
Text input (If present) background, border, height, padding
Code block Background, font, radius, padding
Footer Background, text color, padding
Badge/Pill Background, radius, text style
IDE mockup (If applicable) pane colors, code area style

6. Design Guidelines

Output under the ## Design Guidelines section:

  • Do list: Verifiable specification points for the brand design system (color usage boundaries, font strategy, spacing rules, etc.)
  • Don't list: Explicit prohibitions (don't introduce a second brand color, don't use pure white background, etc.)
  • Responsive breakpoints: If inferable, record breakpoints and layout changes

Extraction Method

  1. Use web_fetch to get the page HTML
  2. Extract style information from:
    • CSS in \x3Cstyle> tags
    • Inline style attributes
    • CSS variables --var-name
    • Computed style patterns (inferred from class names, Tailwind classes, etc.)
    • @font-face declarations
    • Design system files like manifest.json / design tokens
  3. For information not directly accessible from HTML (e.g., hover colors), infer through industry conventions
  4. All color values normalized to HEX format

Notes

  • If precise values cannot be extracted (e.g., design system rendered by JS), mark as estimated with reasoning
  • Font stacks ordered by priority, first entry being the actual brand font
  • Only output components visible on the page; do not construct components not present
  • Information not accessible through static HTML (e.g., animation timings) should be marked as out of scope
安全使用建议
Install only if you are comfortable with the agent fetching the website URL you provide and creating a `{domain}_design.md` file. Avoid giving it private internal URLs or sensitive pages unless you intend that content to be fetched and summarized locally.
能力评估
Purpose & Capability
The requested crawl, style extraction, and `{domain}_design.md` output are coherent with the stated UI design-system scanning purpose.
Instruction Scope
Instructions are scoped to analyzing visible website design attributes and generating structured Markdown; there are no hidden role changes, credential requests, destructive actions, or unrelated data access.
Install Mechanism
The package contains only `SKILL.md`; metadata, static scan, dependency scan, and VirusTotal telemetry show no executable scripts, dependencies, or install-time behavior.
Credentials
The skill requires outbound network access to a user-supplied URL and writes a derived local file, which is expected for this purpose but should be understood by the user.
Persistence & Privilege
Persistence is limited to a named Markdown output file; there is no background worker, privilege escalation, credential/session handling, or ongoing storage behavior.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install ui-scanner
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /ui-scanner 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
- Initial release of UI Scanner: automatically analyzes and extracts a complete visual design system from a website URL. - Outputs a detailed, structured Markdown file using standardized YAML frontmatter and organized design specifications. - Extracts and documents brand color system, typography scales, border radii, spacing values, and major component styles. - Provides high-level design style analysis and usage guidelines (do/don't lists; responsive breakpoints if detectable). - Designed for UI replication, design audits, or reverse engineering of web design systems.
元数据
Slug ui-scanner
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

UI Scanner — Website Design System Extraction 是什么?

Given a website URL, crawl and analyze its visual design system — identify design style, color system, typography, component styles, and UI patterns. Outputs... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 64 次。

如何安装 UI Scanner — Website Design System Extraction?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install ui-scanner」即可一键安装,无需额外配置。

UI Scanner — Website Design System Extraction 是免费的吗?

是的,UI Scanner — Website Design System Extraction 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

UI Scanner — Website Design System Extraction 支持哪些平台?

UI Scanner — Website Design System Extraction 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 UI Scanner — Website Design System Extraction?

由 OpenLark(@openlark)开发并维护,当前版本 v1.0.0。

💬 留言讨论