Token导航 LogoToken导航TokenDH.com
研究检索操作浏览器github未标认证来源可访问许可证需确认审计未展示

ux%3astyle-synthesizeux%3astyle 合成

Agent Skill

用于辅助界面设计、视觉规范、排版、配色、布局和交互体验优化。它适合让 Agent 根据产品场景整理页面结构、生成 UI 方案、检查视觉一致性或改进组件层级。使用时需要结合现有品牌、设计系统和用户任务,不应只堆装饰元素;涉及真实页面改动时,应通过截图或浏览器预览检查文本溢出、对齐和响应式表现。

总安装

261

周安装

11

GitHub Stars

公开资料未说明

下载量

92
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:ux%3astyle-synthesize(ux%3astyle 合成)
来源仓库:https://github.com/cloudvoyant/codevoyant
仓库路径:skills/ux%3Astyle-synthesize
安装命令:
npx skills add https://github.com/cloudvoyant/codevoyant --skill ux:style-synthesize
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/cloudvoyant/codevoyant --skill ux:style-synthesize

简介

用于辅助界面设计、视觉规范、排版、配色、布局和交互体验优化。它适合让 Agent 根据产品场景整理页面结构、生成 UI 方案、检查视觉一致性或改进组件层级。

  • 使用时需要结合现有品牌、设计系统和用户任务,不应只堆装饰元素;涉及真实页面改动时,应通过截图或浏览器预览检查文本溢出、对齐和响应式表现。
  • 适用于研究检索类任务中的风格分析与合成。
  • 需人工复核生成的设计方案可行性。
  • ux%3astyle-synthesize 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Compatibility: The context: fork frontmatter field is Claude Code-specific — on OpenCode and VS Code Copilot it is ignored and the skill runs inline using the current model. Core functionality is preserved on all platforms.

ux:style-synthesize

Visit a live URL, screenshot across breakpoints via browser automation, and synthesize the visual style into a Markdown report and CSS custom-property theme file.

Step 0: Parse Args

Extract from arguments:

  • SOURCE_URL = first non-flag argument that starts with http. If absent, ask: "What URL should we analyze?"
  • NAME = value after --name flag. If absent, derive from URL domain (e.g. linear.app -> linear, www.stripe.com -> stripe)
  • SOURCE_SLUG = slugified NAME (see references/utils.md for slug derivation)
  • OUTPUT_DIR = docs/ux/style-research/{SOURCE_SLUG}/
  • BG_MODE = true if --bg is present
  • SILENT = true if --silent is present

Create the output directory:

mkdir -p {OUTPUT_DIR}/screenshots

Step 1: Open the Site

Use mcp__claude-in-chrome__tabs_context_mcp to check current tabs.

Open a new tab:

mcp__claude-in-chrome__tabs_create_mcp

Navigate to SOURCE_URL:

mcp__claude-in-chrome__navigate: { url: SOURCE_URL, tabId: TAB_ID }

Wait for the page to load (2-3 seconds).

Step 2: Screenshot Across Breakpoints

For each breakpoint in [375, 768, 1440]:

  1. Resize the browser window (see references/utils.md for resize-before-screenshot pattern):
mcp__claude-in-chrome__resize_window: { width: {breakpoint}, height: 900, tabId: TAB_ID }
  1. Scroll to top and wait 1 second for reflow.
  2. Take a screenshot and save to {OUTPUT_DIR}/screenshots/{breakpoint}w.png:
mcp__claude-in-chrome__computer: { action: "screenshot", tabId: TAB_ID }

Save the screenshot image. Note the breakpoint width in your analysis.

  1. Read the page text for content context:
mcp__claude-in-chrome__get_page_text: { tabId: TAB_ID }

Capture visual observations for each breakpoint: layout changes, element visibility, spacing shifts, font size adjustments.

Step 3: Analyze the Design

Based on screenshots and page content, extract:

Color palette:

  • Background color(s) -- primary bg, card bg, sidebar bg
  • Text color(s) -- primary, secondary, muted
  • Brand/accent color(s) -- primary action, links, highlights
  • Border and divider colors
  • Status colors if present (success, error, warning)

Typography:

  • Font families (heading, body, mono if present)
  • Heading scale (H1-H4 approximate sizes and weights)
  • Body text size and line height
  • Letter spacing patterns (e.g. uppercase tracking on labels)

Spacing and layout:

  • Container max-width and horizontal padding
  • Section vertical rhythm (gap between major sections)
  • Card padding, border-radius
  • Grid or flex patterns used

Component patterns:

  • Button styles (filled, outlined, ghost) -- shape, size, shadow
  • Form inputs -- border style, focus ring, label position
  • Navigation -- sticky, transparent, blurred? mobile drawer or hamburger?
  • Cards -- flat, raised, bordered?
  • Tags / badges -- shape, font size, background treatment

Motion / interaction:

  • Hover states (color shifts, underlines, scale)
  • Transitions (duration, easing)
  • Any animations observed

Overall aesthetic:

  • Adjectives that best describe the style (e.g. "minimal, dark, precise, monospace-accented")
  • Key differentiators vs. generic SaaS design

Step 4: Write Style Report

Write {OUTPUT_DIR}/style-report.md using references/style-report-template.md.

Fill in all sections with the observations from Step 3. Include direct observations ("the heading font appears to be DM Sans at ~32px semibold"), not vague statements ("nice typography").

Step 5: Generate CSS Theme

Write {OUTPUT_DIR}/theme.css using references/css-theme-template.md.

Map observations to CSS custom properties:

  • If exact values are identifiable (e.g. #0f0f0f from screenshot color picker context), use them
  • If approximate, use the closest reasonable value and note it with a comment
  • Include both light and dark mode if the site supports it (check via a prefers-color-scheme media query wrapper)

Step 6: Synthesize Summary

Append a one-paragraph "How to apply this style" summary at the end of style-report.md:

  • What Tailwind config changes would replicate this style
  • Which shadcn-svelte theme variables to override
  • Any Google Fonts to load

Step 7: Report + Notify

Print the result summary:

Style synthesis complete: {OUTPUT_DIR}
  style-report.md -- textual analysis
  theme.css -- CSS custom properties

To apply: copy theme.css into your project's app.css and adjust shadcn-svelte variables.

If BG_MODE=true and SILENT=false, send a desktop notification (see references/utils.md):

npx @codevoyant/agent-kit notify --title "ux:style-synthesize complete" --message "Style for {SOURCE_SLUG} synthesized to {OUTPUT_DIR}"

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

33.59%
按下载量换算31

Claude

32.45%
按下载量换算30

Cursor

19.3%
按下载量换算18

Gemini CLI

9.89%
按下载量换算9

安全审计

暂无安全审计结果可展示。

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills