Token导航 LogoToken导航TokenDH.com
前端设计操作浏览器github未标认证来源可访问clear审计通过

cli-ux-designerCLI UX designer 前端

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

1,187

周安装

48

GitHub Stars

11

下载量

372
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:cli-ux-designer(CLI UX designer 前端)
来源仓库:https://github.com/sjungling/claude-plugins
仓库路径:skills/cli-ux-designer
安装命令:
npx skills add https://github.com/sjungling/claude-plugins --skill cli-ux-designer
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/sjungling/claude-plugins --skill cli-ux-designer

简介

cli-ux-designer 提供 CLI 界面设计专家咨询,应用合理默认值与易覆盖原则。

  • 适合优化标志命名、错误消息结构与交互式提示体验。
  • 强调一致性、可发现性与可组合性,避免过度装饰。
  • 安装前应确认项目遵循设计系统约定,避免风格冲突。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

CLI Design Guide

Expert CLI design consultant specializing in creating exceptional command-line interfaces. Design, review, and improve CLI tools by applying comprehensive design principles and patterns.

When NOT to Use This Skill

Do not use this skill for:

  • GUI/web interface design
  • Backend API design (unless CLI tool interacts with it)
  • General UX design outside command-line contexts
  • Programming language design

Core Expertise

Core design principles to apply:

1. Reasonable Defaults, Easy Overrides

  • Optimize for common use cases while providing customization options
  • Use flags to modify default behaviors
  • Consider what most users need most often

2. Maintain Brand Consistency

  • Use platform-specific language and terminology
  • Mirror web interface patterns where appropriate
  • Apply consistent visual styling (colors, states, syntax)
  • Use sentence case, not title case

3. Reduce Cognitive Load

  • Include confirmation steps for risky operations
  • Provide clear headers for context
  • Maintain consistent command patterns
  • Anticipate user mistakes and next actions
  • Design for accessibility

4. Terminal-First with Web Integration

  • Keep users in terminal when possible
  • Provide easy paths to web interface when needed
  • Include --web flags for browser actions
  • Output relevant URLs after operations

Command Structure Expertise

Ensure commands follow this consistent pattern:

tool<command><subcommand>[value][flags][value]
cliissueview234--web-
cliprcreate---title"Title"
clirepoforkorg/repo--clonefalse

Components:

  • Command: The object to interact with
  • Subcommand: The action to take on that object
  • Flag: Modifiers with long version (--state) and often shorthand (-s)
  • Values: IDs, owner/repo pairs, URLs, branch names, file names

Language Guidelines:

  • Use unambiguous language that can't be confused
  • Use shorter phrases when possible and appropriate
  • Use flags for modifiers of actions, avoid making modifiers their own commands
  • Use understood shorthands to save characters

Decision Frameworks

Use these when making CLI design choices:

Flag vs. Subcommand:

  • Flag: modifies HOW a command runs (--verbose, --format json, --dry-run)
  • Subcommand: defines WHAT action to take (issue create, pr merge)
  • Rule: if it changes the action, it's a subcommand. If it changes the behavior, it's a flag.

Interactive vs. Non-interactive:

  • Default to interactive when: user is exploring, multiple choices needed, destructive action requires confirmation
  • Default to non-interactive when: command is commonly scripted, output is piped, CI/CD context
  • Always: provide --yes/-y to skip confirmations, --no-input to disable all prompts

Output Format:

  • Human-readable (default): colors, tables, summaries when stdout is a TTY
  • Machine-readable (piped): no colors, tab-delimited or JSON when stdout is not a TTY
  • Explicit: --format json|table|csv flag to override detection

Error Handling:

  • Exit code 0: success
  • Exit code 1: general error
  • Exit code 2: usage error (wrong flags/args)
  • Always: error message to stderr, suggested fix when possible

Visual Design System Knowledge

Typography

  • Assume monospace fonts
  • Use bold for emphasis and repository names
  • Create hierarchy with spacing and weight
  • No italics (unreliable support)

Color Usage

Apply the 8 basic ANSI colors:

  • Green: Success, open states
  • Red: Failure, closed states
  • Yellow: Warnings, draft states
  • Blue: Information, links
  • Cyan: Branch names, special identifiers
  • Magenta: Special highlights
  • Gray: Secondary information, labels
  • White/Default: Primary text

Guidelines:

  • Only enhance meaning, never communicate meaning solely through color
  • Consider users can customize terminal colors
  • Some terminals don't support 256-color sequences reliably

For complete ANSI color codes and escape sequences, see ./references/ansi-color-reference.md.

Iconography

Use Unicode symbols consistently:

  • Success
  • Failure
  • ! Alert
  • - Neutral
  • + Changes requested

Consider varying Unicode font support across systems.

For a comprehensive list of CLI-friendly Unicode symbols, see ./references/unicode-symbols.md.

Component Pattern Expertise

Lists

  • Use tabular format with headers
  • Show state through color
  • Include relevant contextual information

For a complete list view example, see ./references/examples/list-view-example.txt.

Detail Views

  • Show comprehensive information
  • Indent body content
  • Include URLs at bottom

Prompts

  • Yes/No: Default in caps, for confirmations
  • Short text: Single-line input with autocomplete
  • Long text: Multi-line with editor option
  • Radio select: Choose one option
  • Multi-select: Choose multiple options
  • Always provide flag alternatives to prompts

For an interactive prompt example, see ./references/examples/interactive-prompt-example.txt.

Help Pages

Required sections: Usage, Core commands, Flags, Learn more, Inherited flags Optional sections: Additional commands, Examples, Arguments, Feedback

For a complete help text example, see ./references/examples/help-text-example.txt.

Syntax Conventions

  • <required-args> in angle brackets
  • [optional-args] in square brackets
  • {mutually-exclusive} in braces
  • repeatable... with ellipsis
  • Use dash-case for multi-word variables

Anti-patterns

Avoid these common CLI design mistakes:

Anti-patternBetter Approach
Deeply nested subcommands (tool group sub action)Max 2 levels: tool command [flags]
Inconsistent flag naming (--no-color vs --disable-colors)Pick one convention and apply everywhere
Interactive prompts with no flag alternativesEvery prompt must have a --flag equivalent
Cryptic error messages ("Error: 1")Include what went wrong, why, and how to fix
Silent failures (exit 0 on error)Non-zero exit codes for failures, stderr for errors
Missing --help on subcommandsEvery command level should have help
Mixing stdout data with status messagesData to stdout, progress/status to stderr

Technical Considerations

Script Automation Support

  • Provide flags for all interactive elements
  • Output machine-readable formats when piped
  • Use tabs as delimiters for structured data
  • Remove colors/formatting in non-terminal output
  • Include exact timestamps and full data

Accessibility

  • Use punctuation for screen reader pauses
  • Don't rely solely on color for meaning
  • Support high contrast and custom themes
  • Design for cognitive accessibility

Success Criteria

Recommendations are successful when:

  • Commands follow consistent patterns across the tool
  • Help text is clear with useful examples
  • Visual hierarchy guides users naturally
  • Both interactive and scriptable use cases work
  • Accessibility requirements are met

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

30.43%
按下载量换算113

windsurf

23.27%
按下载量换算87

trae

15.95%
按下载量换算59

OpenCode

11.47%
按下载量换算43

Codex

8.07%
按下载量换算30

Antigravity

3.24%
按下载量换算12

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。

来源信息

继续浏览同类 Skills