Token导航 LogoToken导航TokenDH.com
前端设计需要联网github未标认证来源可访问clear审计通过

ui-template-library用户界面模板库

Agent Skill

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。

总安装

349

周安装

15

GitHub Stars

182

下载量

122
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

3

许可证

MIT

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:ui-template-library(用户界面模板库)
来源仓库:https://github.com/glittercowboy/plugin-freedom-system
仓库路径:skills/ui-template-library
安装命令:
npx skills add https://github.com/glittercowboy/plugin-freedom-system --skill ui-template-library
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-template-library

简介

用于提供可复用的前端界面模板和组件集合。

  • 适合快速生成表单、卡片、导航等常见 UI 区块。
  • 包含多种布局和主题风格的代码示例。ui-template-library 属于前端设计类 Skill,可作为该场景下的辅助能力补充。
  • 使用时需适配项目技术栈和设计语言。
  • 适用于加速开发流程和保持界面风格统一。

SKILL.md

ui-template-library Skill

Purpose: Capture and apply aesthetic "vibes" across plugins using structured prose descriptions rather than rigid specifications.


Operations

You MUST determine which operation the user is requesting, then execute ONLY that operation.

Operation Routing (match user request to keywords):

  • save/capture/extract → save_aesthetic (Operation 1)
  • apply/use/generate → apply_aesthetic (Operation 2)
  • list/show/browse → list_aesthetics (Operation 3)
  • delete/remove → delete_aesthetic (Operation 4)
  • update/modify/refine → update_aesthetic (Operation 5)
  1. save_aesthetic - Analyze mockup HTML and generate structured prose aesthetic.md
  2. apply_aesthetic - Interpret aesthetic.md to generate new mockup with appropriate layout
  3. list_aesthetics - Show all saved aesthetics with prose summaries
  4. delete_aesthetic - Remove aesthetic from library
  5. update_aesthetic - Refine aesthetic.md from improved mockup

CRITICAL: Checkpoint Verification Protocol

Before presenting decision menu in ANY operation:

ALL steps must complete successfully. If ANY step fails: HALT, report specific error, DO NOT proceed to decision menu.

Verification checklist (run before showing menu):

  • All required files created successfully
  • File contents are valid (aesthetic.md not empty, metadata.json valid JSON)
  • manifest.json updated successfully
  • Git commit completed (if applicable)
  • No error states from previous steps

Error handling:

  • Steps 1-6 fail → Report error, do not create partial aesthetic directory
  • Step 7 fails (manifest) → Aesthetic files exist but not registered, offer retry
  • Step 7.5 fails (git) → Non-fatal, warn "Aesthetic saved but not committed"

Note: Always use inline numbered menus (NOT AskUserQuestion tool) to maintain consistency with system-wide Checkpoint Protocol.


Directory Structure

.claude/aesthetics/
├── manifest.json                    # Registry of all aesthetics
└── [aesthetic-id]/
    ├── aesthetic.md                 # Structured prose description (THE SOURCE OF TRUTH)
    ├── preview.html                 # Visual reference (original mockup)
    └── metadata.json                # Name, description, source plugin, tags

Key change from previous system:

  • aesthetic.yaml → aesthetic.md: YAML specs replaced with structured prose descriptions
  • Format is consistent and parseable, but content is interpretable concepts, not rigid values

<critical_sequence> Initialize.claude/aesthetics/ directory if missing (mkdir -p) Validate mockup file exists and is readable Read mockup HTML file into memory (Read tool) Extract visual patterns using pattern-extraction.md strategies Generate prose descriptions using prose-generation.md guidelines Write aesthetic.md following aesthetic-template.md structure exactly: same section headers in same order, fill all sections with prose (no placeholders) (load template from assets/aesthetic-template.md, if missing: report critical error) Copy preview.html to aesthetic directory Generate metadata.json with inferred tags Update manifest.json using Read → modify → Write pattern (if missing: initialize from assets/manifest-init.json) Commit aesthetic to git using conventional format Present confirmation + decision menu </critical_sequence>

<state_requirement> See "CRITICAL: Checkpoint Verification Protocol" section for complete verification requirements. </state_requirement>

Operation 1: save_aesthetic

See: references/save-operation.md for complete 8-step workflow.


<critical_sequence> Load aesthetic.md and metadata.json (Read tool) Read target plugin parameter-spec.md or creative-brief.md Interpret aesthetic prose using aesthetic-interpretation.md strategies Choose layout based on parameter count (layout-generation.md) Generate HTML/CSS applying aesthetic to chosen layout Save generated mockup to target plugin mockups directory Update aesthetic metadata.json usedInPlugins array Present decision menu as inline numbered list, NOT AskUserQuestion tool (rationale: consistent with system-wide Checkpoint Protocol - see CLAUDE.md) </critical_sequence>

<decision_gate wait_required="true"> Present decision menu with preview/continue options MUST wait for user response </decision_gate>

Operation 2: apply_aesthetic

See: references/apply-operation.md for complete 8-step workflow.


<decision_gate wait_required="true"> Wait for user to select action from menu </decision_gate>

Operation 3: list_aesthetics

Display all saved aesthetics in table format with preview paths. See: references/list-operation.md for complete 4-step workflow.


Operation 4: delete_aesthetic

Remove aesthetic from library with confirmation. See: references/delete-operation.md for complete 4-step workflow.


Operation 5: update_aesthetic

Refine aesthetic.md from improved mockup while preserving user edits. See: references/update-operation.md for complete 7-step workflow.


API Contract

This skill provides operations for managing aesthetic templates. Called by ui-mockup skill.

save_aesthetic(mockup_path, plugin_name, aesthetic_name?):

  • Input: Absolute path to mockup HTML, plugin name, optional aesthetic name
  • Output: aesthetic_id (string), aesthetic_path (absolute path)
  • Process: Analyzes mockup, generates aesthetic.md prose, saves to.claude/aesthetics/

apply_aesthetic(aesthetic_id, plugin_name, parameter_spec_path):

  • Input: Aesthetic ID, target plugin name, path to parameter-spec.md
  • Output: mockup_path (absolute path to generated mockup)
  • Process: Loads aesthetic prose, interprets to CSS, generates mockup with appropriate layout

Invocation: Inline invocation from ui-mockup (both skills are lightweight and stateless).


Success Criteria

Save operation successful when:

  • ✅ aesthetic.md generated following exact template structure
  • ✅ All sections filled with interpretable prose (no placeholders)
  • ✅ Example color codes provided as concrete reference
  • ✅ preview.html copied to aesthetic directory
  • ✅ metadata.json created with tags and source info
  • ✅ manifest.json updated with new aesthetic entry
  • ✅ Format is idempotent (same mockup → same structure every time)

Apply operation successful when:

  • ✅ aesthetic.md parsed and interpreted correctly
  • ✅ Appropriate layout chosen for parameter count
  • ✅ Generated mockup reflects aesthetic visual language
  • ✅ Control styling matches prose descriptions
  • ✅ Colors, typography, spacing consistent with aesthetic
  • ✅ WebView constraints enforced (no viewport units, etc.)
  • ✅ Result is recognizably the same aesthetic despite different layout

List operation successful when:

  • ✅ All aesthetics shown in readable table
  • ✅ Vibe summaries extracted from prose
  • ✅ Preview paths provided for visual reference
  • ✅ Decision menu presented for next action

Implementation Notes

Technical details about parsing, quality control, and constraints. See: references/implementation-notes.md for complete technical documentation.


Reference Documentation

  • Pattern extraction: references/pattern-extraction.md - String/regex strategies for extracting visual patterns from HTML
  • Prose generation: references/prose-generation.md - Guidelines for transforming patterns into interpretable prose
  • Aesthetic interpretation: references/aesthetic-interpretation.md - Strategies for parsing prose and generating CSS/HTML
  • Layout generation: references/layout-generation.md - Layout decision trees and control placement strategies

Template Assets

  • Aesthetic template: assets/aesthetic-template.md - Complete structured prose template (THE FORMAT SPEC)
  • Metadata template: assets/metadata-template.json - Metadata JSON structure
  • Manifest init: assets/manifest-init.json - Empty manifest structure for initialization

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

04

需要参考平台分布和安装热度时

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

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

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

Claude Code

31.4%
按下载量换算38

windsurf

24.15%
按下载量换算29

OpenCode

16.74%
按下载量换算20

Codex

13.65%
按下载量换算17

Antigravity

7.22%
按下载量换算9

Gemini CLI

3.4%
按下载量换算4

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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。

来源信息

继续浏览同类 Skills