Token导航 LogoToken导航

Agent Skill · 前端设计

ui-mockup(用户界面样机)

用于辅助前端页面、组件、样式和交互逻辑的开发与维护。它适合让 Agent 生成或审查 React、Next.js、Vue、Tailwind、CSS 等相关代码,整理组件结构,或定位布局和性能问题。使用时需要结合项目现有设计系统、路由和构建方式,避免只生成孤立片段;涉及页面改动时,应配合本地预览和构建检查确认视觉效果。

ui-mockup 是前端设计类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为操作浏览器;许可证为 MIT,第三方扫描已通过,按安装量在前端设计类的 10,216 个 Skill 里排第 6,045。

总安装 563周安装 23GitHub Stars 182下载 180核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

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:ui-mockup(用户界面样机)
来源仓库:https://github.com/glittercowboy/plugin-freedom-system
仓库路径:skills/ui-mockup
安装命令:
npx skills add https://github.com/glittercowboy/plugin-freedom-system --skill ui-mockup
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/glittercowboy/plugin-freedom-system --skill ui-mockup

这个 Skill 是做什么的

前端设计类 Skill 关注页面结构、视觉规范和组件写法,输出通常是可运行的界面代码。

用于辅助前端页面、组件、样式和交互逻辑的开发与维护。适合生成或审查 React、Next.js、Vue、Tailwind、CSS 等相关代码,整理组件结构或定位布局和性能问题。使用时需结合项目现有设计系统、路由和构建方式,避免生成孤立片段;涉及页面改动时应配合本地预览和构建检查确认视觉效果。

安装方式:github,可通过 npx skills add 命令从指定仓库添加。适用于 Codex、Claude、Cursor、Gemini CLI 等宿主环境。

来源仓库内路径:skills/ui-mockup

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/glittercowboy/plugin-freedom-system --skill ui…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要操作浏览器,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

ui-mockup 怎么安装?

在终端执行 npx skills add https://github.com/glittercowboy/plugin-freedom-system --skill ui-mockup。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

ui-mockup 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

ui-mockup 需要什么权限?

权限级别是操作浏览器。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。

ui-mockup 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

ui-mockup 用的人多吗?

公开数据是累计安装 563 次,近一周安装 23 次,下载 180 次,GitHub 182 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

ui-mockup 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理前端设计类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

前端设计类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
frontend-design

前端设计 · Codex、Claude

8,823,360126,072只读MIT通过
vercel-composition-patterns

Vercel 组合模式 · Codex、Claude

3,858,81625,893需要联网MIT通过
vercel-react-native-skills

Vercel React native skills 前端 · Codex、Claude

2,468,66425,878只读MIT通过
adapt

adapt 搜索 · Codex、Claude

2,027,04023,125操作浏览器未明确通过
bolder

bolder 分析 · Codex、Claude

2,004,28823,130只读未明确通过
ui-mockup

用户界面样机 · 本页

563182操作浏览器MIT通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

<quick_start>

  1. Check for aesthetic library (Phase 0)
  2. Gather requirements through tiered questions (Phases 1-3)
  3. Dispatch ui-design-agent for mockup generation
  4. Iterate until user approves
  5. Generate implementation files </quick_start>

<context_detection> <standalone_mode>

  • No .continue-here.md file present
  • Generates mockups independently
  • Skips state updates </standalone_mode>

<workflow_mode>

  • File plugins/[PluginName]/.continue-here.md exists with current_stage field
  • Updates workflow state after each phase </workflow_mode>

Check for .continue-here.md existence to determine mode. If present, update state files. If absent, skip state updates. </context_detection>

How would you like to start the UI design?

  1. Start from aesthetic template - Apply saved visual system
  2. Start from scratch - Create custom design
  3. List all aesthetics - Browse library before deciding

Choose (1-3): _

See references/aesthetic-integration.md for complete integration details.

Include in invocation prompt:

  • All gathered requirements (layout, controls, colors, etc.)
  • Quality expectation: "Design must look like commercial $50-200 audio plugin - intentional decisions, not defaults"

See references/phase-details.md for invocation summary. See references/delegation-protocols.md for complete protocol.

Files generated:

  • v[N]-ui.yaml (design specification)
  • v[N]-ui-test.html (browser-testable mockup)

What would you like to do?

  1. Iterate - Refine design, adjust layout
  2. Finalize - Validate alignment and complete mockup
  3. Save as template - Add to aesthetic library for reuse
  4. Other

Choose (1-4): _

<iteration_guidance> When collecting feedback for Option 1 (Iterate):

If user provides specific changes ("make it vertical", "add a meter", "change colors to blue"):

  • Pass these as explicit requirements to new ui-design-agent instance

If user provides vague improvement requests ("make it better", "improve it", "polish it"):

  • Prompt for specifics: "What aspect should I focus on? (layout, colors, spacing, controls)"
  • If user says "everything" or "overall quality": Pass instruction to refine existing elements (spacing, color harmony, control styling) rather than add new elements </iteration_guidance>

See references/decision-menus.md#phase-5-5-design-decision-menu for detailed routing.

See references/phase-b-enforcement.md for guard implementation.

<orchestration_protocol> <delegation_rules> This skill NEVER generates mockup files directly. ALL file generation delegated to subagents.

See references/delegation-protocols.md for enforcement details. </delegation_rules>

<state_management> Subagents update .continue-here.md with their phase results. Orchestrator verifies stateUpdated flag in JSON report.

After subagent returns stateUpdated: true, verify actual state contents match expected values (not just boolean flag).

Read .continue-here.md, parse YAML, check specific fields match JSON report values. If mismatch, present state recovery menu.

See references/state-tracking.md for complete state schema and verification protocol. </state_management>

<iteration_protocol> Each iteration runs in fresh agent context. User chooses "Iterate" then orchestrator collects feedback and invokes NEW ui-design-agent with incremented version. Fresh context prevents context window bloat during iterative design. </iteration_protocol>

<error_handling>

  • Agent failures: Present error menu (retry/manual fix/debug/cancel).
  • Validation failures: Agent returns validationPassed: false - present error menu.
  • State update failures: Agent returns stateUpdated: false - present state recovery menu (verify/manual update/continue anyway). </error_handling> </orchestration_protocol>

<versioning_strategy> Pattern: v1, v2, v3... Each version saved separately.

Purpose: Explore layouts without losing work, A/B test designs, keep history for rollback.

File naming: All 7 files prefixed with version (e.g., v2-ui.html, v2-PluginEditor.h).

Implementation: Latest version used for Stage 3 (GUI) unless user specifies different version.

See references/versioning.md for file management details. </versioning_strategy>

<success_criteria> <design_phase>

  • YAML spec generated matching user requirements
  • Browser test HTML works (interactive controls, parameter messages)
  • Visual quality meets commercial standard (intentional design, not defaults)
  • Design files committed to git
  • .continue-here.md updated with version (if workflow mode)
  • User presented with Phase 5.5 decision menu
  • Design approved OR user iterates with refinements </design_phase>

<implementation_phase>

  • All 7 files generated and saved to .ideas/mockups/
  • Production HTML complete (no placeholders)
  • C++ boilerplate matches YAML structure (correct parameter bindings)
  • parameter-spec.md generated and locked (v1 only)
  • Implementation files committed to git
  • .continue-here.md updated with finalization status (if workflow mode) </implementation_phase> </success_criteria>

<integration_points> <invoked_by>

  • /dream command - After creative brief, before implementation
  • plugin-workflow skill - During Stage 0 (UI design phase)
  • plugin-improve skill - When redesigning existing plugin UI
  • Natural language: "Design UI for [PluginName]", "Create mockup for compressor" </invoked_by>

<always_invokes>

  • ui-design-agent subagent (Phase 4-5.45) - REQUIRED for design iteration
  • ui-finalization-agent subagent (Phase 6-10.5) - REQUIRED for implementation files </always_invokes>

<also_invokes>

  • ui-template-library skill (if user saves aesthetic) </also_invokes>

<reference_documentation> Progressive disclosure - load references when reaching specific phases:

  • Phase 0: references/aesthetic-integration.md - Aesthetic library integration
  • Phase 1: references/context-extraction.md - Creative brief extraction
  • Phase 2-3: references/design-questions.md - Question templates and tiering
  • Phase 2.5: references/layout-validation.md - Dimension calculation formulas
  • Phase 4-5.45: references/delegation-protocols.md - Subagent invocation
  • Phase 5.5: references/decision-menus.md - Menu format and routing
  • Phase 5.6: references/state-tracking.md - Brief sync protocol
  • Phase 6-10: references/phase-b-enforcement.md - Phase B guard
  • Anti-patterns: references/common-pitfalls.md - What to avoid

Technical details:

  • references/html-generation.md - Production HTML rules
  • references/browser-testing.md - Browser test workflow
  • references/cmake-configuration.md - WebView build settings
  • references/cpp-boilerplate-generation.md - C++ template generation
  • references/ui-design-rules.md - Design constraints and patterns
  • references/versioning.md - File management </reference_documentation>

<template_assets>

  • assets/ui-yaml-template.yaml - YAML structure
  • assets/webview-boilerplate.md - C++ integration templates
  • assets/integration-checklist-template.md - Integration guide
  • assets/parameter-spec-template.md - Parameter specification format </template_assets>

平台分布

Claude Code

27.33%
按下载量换算49

OpenCode

21.16%
按下载量换算38

windsurf

19.26%
按下载量换算35

Codex

14.5%
按下载量换算26

Antigravity

8.05%
按下载量换算14

Gemini CLI

3.88%
按下载量换算7

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

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。

继续浏览同类 Skills