Token导航 LogoToken导航TokenDH.com
开发规范执行命令github未标认证来源可访问许可证需确认审计提醒

define-design-guidelines定义设计指南

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

242

周安装

10

GitHub Stars

11

下载量

79
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:define-design-guidelines(定义设计指南)
来源仓库:https://github.com/doodledood/claude-code-plugins
仓库路径:skills/define-design-guidelines
安装命令:
npx skills add https://github.com/doodledood/claude-code-plugins --skill define-design-guidelines
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/doodledood/claude-code-plugins --skill define-design-guidelines

简介

制定界面视觉、交互模式和动效规范的完整设计体系。

  • 适用于移动端、Web 和跨平台产品的体验一致性保障。
  • 需结合用户画像和使用场景确定信息层级结构。
  • 输出包含色彩系统、字体规范和组件行为定义。
  • define-design-guidelines 属于开发规范类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Design Guidelines Skill

Create the DESIGN_GUIDELINES.md document that defines HOW to design interfaces for your customer. This document drives all UI/UX: components, layouts, animations, colors, typography—everything visual.

Prerequisite: CUSTOMER.md must exist. Design without customer definition is just aesthetic preference. The interface must resonate with WHO you're building for.

Overview

This skill supports both creating new design guidelines and refining existing ones.

This skill guides you through: 0. Prerequisite Check - Verify CUSTOMER.md exists; stop if not

  1. Deep Analysis - Launch design-research agent to understand ideal design for the customer
  2. Discovery - Confirm/refine design direction with targeted questions
  3. Generate Document - Create DESIGN_GUIDELINES.md with full design system
  4. Automatic Alignment Audit - Opus agent verifies alignment with CUSTOMER.md/BRAND_GUIDELINES.md, fixes issues, repeats until perfect
  5. Finalization - Add version history once audit passes

Core Philosophy: Anti-AI-Slop

CRITICAL: This skill bakes in the frontend-design skill's principles. Every design guideline must avoid generic AI aesthetics:

What to AVOID (AI Slop)

  • Generic fonts: Inter, Roboto, Arial, system fonts, Space Grotesk
  • Cliché colors: Purple gradients on white, generic blue CTAs, safe gray palettes
  • Predictable layouts: Cookie-cutter grids, template-looking compositions
  • Safe choices: Border-radius everywhere, subtle animations, inoffensive everything
  • Generic components: Bootstrap/MUI defaults without personality

What to EMBRACE

  • Bold aesthetic commitment: Pick an extreme and execute with precision
  • Distinctive typography: Characterful fonts that match the product personality
  • Intentional color: Dominant colors with sharp accents, not timid even distribution
  • Spatial creativity: Asymmetry, overlap, diagonal flow, grid-breaking elements
  • Atmospheric details: Textures, gradients, shadows, effects that create depth

The test: Would someone mistake this for a generic template? If yes, it's wrong.

Workflow

Phase 0: Prerequisite Check

CRITICAL: Before anything else, check for CUSTOMER.md:

  1. Use Glob to search for **/CUSTOMER.md in the current directory
  2. If NOT found: Stop immediately and inform the user:
"I can't create design guidelines without knowing WHO you're designing for.

Please create your CUSTOMER.md first using /define-customer.

Design without customer definition is just aesthetic preference—it won't resonate with anyone specific."

Do NOT proceed. End the workflow here.

  1. If found: Read the CUSTOMER.md and extract key context:

- ICP definition (who they are) - What they value (speed? precision? fun? simplicity?) - Behavioral traits (patient? impatient? technical? casual?) - Anti-personas (who they're NOT) - Any visual/experience hints

IMPORTANT - Pre-fill Recommendations: Use CUSTOMER.md to infer recommended options for all questions:

If CUSTOMER.md says...Recommend...
ICP values "speed", "efficiency", "no patience"Terminal/utilitarian aesthetic, fast animations, dense UI
ICP is technical (developers, engineers)Monospace typography, dark theme, information-dense
ICP values "data", "statistics", "precision"Data terminal aesthetic, clinical colors, sharp geometry
ICP is "fun-seekers", "casual players", "beginners"Playful/soft aesthetic, rounded corners, inviting colors
ICP is "professionals", "executives"Refined/luxury aesthetic, premium typography, restrained palette
ICP values "creativity", "expression"Bold/maximalist aesthetic, unexpected layouts, strong personality
Anti-persona is "corporate" or "enterprise"Avoid generic SaaS look, embrace distinctive character

The goal: User should be able to accept all recommended defaults and get a design system that resonates with their ICP.

Then check for existing DESIGN_GUIDELINES.md:

header: "Existing Design Guidelines Found"
question: "I found existing DESIGN_GUIDELINES.md. What would you like to do?"
options:
  - "Refine it - update based on new insights"
  - "Start fresh - create new design guidelines"
  - "Review it - just read through what's there"

Phase 1: Deep Analysis

BEFORE asking any questions, launch the design-research agent to deeply analyze the customer profile and determine the ideal design direction.

Launch the Design Research Agent:

Launch Task agent (subagent_type: design-research) with prompt:

"Analyze the customer profile to determine ideal UI/UX design direction.

CUSTOMER.md path: [path found in Phase 0]

Provide your full design analysis covering:
1. Customer Design Psychology
2. Recommended Aesthetic Direction
3. Typography Recommendation
4. Color Direction
5. Geometry & Motion
6. Signature Elements
7. Anti-Patterns for This ICP
8. Design Reference Products

Be specific and decisive. This analysis will inform the entire design system."

The agent will:

  1. Read CUSTOMER.md and BRAND_GUIDELINES.md (if exists)
  2. Research industry design patterns and competitors
  3. Provide comprehensive design analysis with specific recommendations

After agent completes, extract the analysis and use it to:

  1. Pre-fill ALL question recommendations with high confidence
  2. Present a summary to the user before discovery questions

Present Analysis Summary:

header: "Design Analysis"
question: "Based on your customer profile, here's my recommended design direction. Does this feel right?"
[Display: Aesthetic direction, typography, theme, key signature elements]
options:
  - "Yes - this direction feels right, let's refine details (Recommended)"
  - "Mostly - good direction but some things feel off"
  - "No - I have a different vision"

If "Yes" or "Mostly", proceed to Phase 2 with agent recommendations as defaults. If "No", ask what's different and adjust recommendations.

Phase 2: Discovery

Use AskUserQuestion for all questions. Put the recommended option FIRST with "(Recommended)" suffix. Use the opus agent's analysis to inform ALL recommendations.

Question 1: Aesthetic Direction

This is the most important question. The entire design system flows from this choice.

Use the opus agent's recommendation as the default. The agent has already analyzed CUSTOMER.md deeply.

header: "Aesthetic Direction"
question: "What aesthetic direction fits your product and customer?"
options:
  - "[Inferred from CUSTOMER.md] (Recommended)"
  - "Data terminal - clinical, sharp, information-dense (Bloomberg, trading apps)"
  - "Brutally minimal - stark, essential, no decoration"
  - "Industrial utilitarian - functional, raw, tool-like"
  - "Luxury/refined - premium, elegant, restrained"
  - "Editorial/magazine - typographic, editorial, sophisticated"
  - "Brutalist/raw - bold, unapologetic, confrontational"
  - "Retro-futuristic - nostalgic tech, synthwave, neon"
  - "Playful/toy-like - fun, colorful, delightful"
  - "Soft/pastel - gentle, approachable, calming"
  - "Art deco/geometric - structured, ornamental, patterns"
  - "Organic/natural - flowing, earthy, warm"

Question 2: Theme Preference

Use the opus agent's color direction analysis.

header: "Theme"
question: "What's your primary theme?"
options:
  - "[Inferred] (Recommended)"
  - "Dark theme - dark backgrounds, light text (more distinctive)"
  - "Light theme - light backgrounds, dark text (more accessible)"
  - "Both - design for both with theme switching"

Question 3: Typography Character

Use the opus agent's typography recommendation.

header: "Typography"
question: "What typographic character fits your brand?"
options:
  - "[Inferred] (Recommended)"
  - "Monospace-forward - technical, precise, data-focused (JetBrains Mono, Fira Code)"
  - "Elegant serif - premium, editorial, sophisticated (Playfair, Cormorant)"
  - "Bold geometric - strong, modern, impactful (Clash Display, Satoshi)"
  - "Rounded/friendly - approachable, soft, inviting (Nunito, Quicksand)"
  - "Editorial mix - display headlines with refined body (custom pairing)"
  - "Clean sans - neutral but NOT generic (Geist, DM Sans - not Inter/Roboto)"

Question 4: Geometry & Shape

Use the opus agent's geometry & motion analysis.

header: "Geometry"
question: "What geometric character defines your UI?"
options:
  - "[Inferred] (Recommended)"
  - "Sharp zero-radius - all corners sharp, no exceptions (precision, clinical)"
  - "Minimal/subtle - small radius (4-6px) for polish without softness"
  - "Rounded/soft - generous radius (8-16px) for friendliness"
  - "Pill shapes - fully rounded buttons/badges for playfulness"
  - "Mixed - sharp containers, rounded interactive elements"

Question 5: Information Density

Use the opus agent's analysis of ICP technical level and patience.

header: "Density"
question: "How dense should information be?"
options:
  - "[Inferred] (Recommended)"
  - "Dense - information-rich, minimal whitespace (power users)"
  - "Balanced - comfortable density with clear hierarchy"
  - "Spacious - generous whitespace, breathing room"
  - "Editorial - dramatic spacing, statement pieces"

Question 6: Motion Philosophy

Use the opus agent's motion philosophy recommendation.

header: "Motion"
question: "What's your animation philosophy?"
options:
  - "[Inferred] (Recommended)"
  - "Instant - <100ms, no unnecessary animation (respects time)"
  - "Functional - fast, purposeful, feedback-focused"
  - "Subtle - refined micro-interactions, polished feel"
  - "Delightful - playful animations, personality-forward"
  - "Dramatic - bold transitions, statement animations"

Question 7: Primary Color Direction

header: "Primary Color"
question: "What color family anchors your palette?"
options:
  - "[Inferred if clear signal] (Recommended)"
  - "Orange/amber - energy, action, warmth"
  - "Blue - trust, calm, professional"
  - "Green - growth, success, nature"
  - "Purple - creativity, premium, unique"
  - "Red/coral - bold, urgent, passionate"
  - "Teal/cyan - modern, tech, fresh"
  - "Neutral - black/white/gray dominant, accent secondary"
  - "Custom - I have specific brand colors"

Question 7b: Brand Colors (If "Custom" selected)

header: "Brand Colors"
question: "What are your brand colors?"
freeText: true
placeholder: "e.g., 'Primary: #F97316 (orange), Secondary: #3B82F6 (blue), Background: #0A0A0B'"

Question 8: Technical Constraints

header: "Tech Stack"
question: "What's your frontend tech stack?"
options:
  - "React + Tailwind (Recommended - most flexible)"
  - "React + CSS-in-JS (styled-components, emotion)"
  - "React + CSS Modules"
  - "Vue + Tailwind"
  - "Vanilla HTML/CSS/JS"
  - "Other framework"
multiSelect: false

Question 9: Signature Elements

header: "Signature"
question: "What should be immediately recognizable about your UI? (Select 2-3)"
options:
  - "[Inferred from aesthetic] (Recommended)"
  - "Zero border-radius everywhere"
  - "Monospace typography for data"
  - "Single dominant accent color"
  - "Heavy use of negative space"
  - "Custom cursor or micro-interactions"
  - "Unique loading states"
  - "Distinctive iconography"
  - "Gradient treatments"
  - "Texture or grain overlays"
  - "Bold asymmetric layouts"
multiSelect: true

Question 10: Product Context (Free text)

header: "Product Context"
question: "Describe your product briefly - what does it do and what's the primary interface?"
freeText: true
placeholder: "e.g., 'Chess analysis tool - main screen shows a chess board with win percentage overlay. Data-heavy stats pages.'"

Question 11: References (Optional, free text)

header: "References"
question: "Any products or websites whose design you admire? (Optional)"
freeText: true
placeholder: "e.g., 'Linear's clean interface, Stripe's documentation, Notion's typography'"

Question 12+: Gap-Filling

After core questions, verify you have clarity on:

  • Aesthetic direction (specific, not vague)
  • Typography approach
  • Color palette direction
  • Geometry decisions
  • Motion philosophy
  • Signature elements

Keep asking until confident enough to generate a distinctive design system.

Phase 3: Generate Document

Based on the design-research agent's analysis and user's confirmed preferences, generate DESIGN_GUIDELINES.md.

The document should include:

  1. Identity - User, problem, aesthetic direction, "We Are / We Are NOT", signature elements, core principles
  2. Design Tokens - Colors (surfaces, text, accent, status, borders), typography (fonts, sizes, weights), spacing, geometry, shadows, animation timings, breakpoints
  3. Voice & Copy - UI tone, bad/good examples, state copy
  4. Components - Cards, buttons, inputs, toasts with specific specs
  5. Layout Patterns - Primary layout philosophy, visual hierarchy
  6. Motion - Philosophy, orchestrated reveals, loading states
  7. Anti-Patterns - AI slop to avoid, brand violations, customer-ignoring mistakes
  8. Ship Checklist - Pre-ship verification items

Incorporate the design-research agent's full analysis into the document—especially the anti-patterns, signature elements, and reference products.

Write DESIGN_GUIDELINES.md to the current working directory.

Phase 4: Automatic Alignment Audit

After generating the document, automatically audit it against CUSTOMER.md and BRAND_GUIDELINES.md to ensure perfect alignment. This is NOT optional—run at least one audit cycle.

Launch the Design Quality Auditor Agent:

Launch Task agent (subagent_type: design-quality-auditor) with prompt:

"Audit DESIGN_GUIDELINES.md for alignment with customer profile and brand guidelines.

Document paths:
- DESIGN_GUIDELINES.md: [path]
- CUSTOMER.md: [path]
- BRAND_GUIDELINES.md: [path] (if exists)

Perform your full audit protocol and report results."

The agent will:

  1. Read all three documents
  2. Check customer alignment, brand alignment, internal consistency, completeness
  3. Report ✅ AUDIT PASSED or ⚠️ ISSUES FOUND with specific fixes

After audit completes:

  1. If AUDIT PASSED: Proceed to finalization
  2. If ISSUES FOUND:

- Apply all suggested fixes to DESIGN_GUIDELINES.md - Run the audit again - Repeat until AUDIT PASSED (max 3 cycles to prevent infinite loops)

Example audit issue and fix:

ISSUE: Customer Alignment - ICP values "zero patience" but motion philosophy
specifies 300ms animations which feels slow for this audience.
FIX: Change base animation duration to 100ms, reserve 300ms only for
celebratory moments like score reveals.

Phase 5: Finalization

Once audit passes, add version history:

---

## Version History

- **v1.0** - [Date] - Initial creation (audit passed)

## Usage

Reference this document for ALL UI work. The test: Would your ICP feel this UI was made for them?

Key Principles

Grounded in Customer

  • Every design choice should resonate with the ICP
  • Reference CUSTOMER.md values when making decisions
  • If the ICP values speed, the UI must be fast
  • If the ICP is technical, the UI can be dense

Anti-AI-Slop is Non-Negotiable

  • Every design system must be distinctive
  • Generic choices are wrong by default
  • If it looks like a template, it fails
  • Bold commitment beats safe mediocrity

Actionable Over Abstract

  • Don't just say "be bold" - specify the border-radius
  • Every guideline needs exact values
  • Ship checklist catches drift

Aesthetic Coherence

  • All tokens must serve the chosen direction
  • Typography, color, motion, geometry align
  • Signature elements appear consistently
  • Deviations are intentional, not accidental

Reduce Cognitive Load

  • ALWAYS use AskUserQuestion tool when available
  • Put recommended option FIRST with "(Recommended)" suffix
  • Pre-fill recommendations from agent analysis - user should be able to accept defaults
  • Present multi-choice questions to minimize typing

Output Location

Write DESIGN_GUIDELINES.md to the current working directory (or user-specified path).

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

能力 4

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

32.88%
按下载量换算26

Codex

31.88%
按下载量换算25

Cursor

19.57%
按下载量换算15

Gemini CLI

10.18%
按下载量换算8

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/doodledood/claude-code-plugins --skill define-design-guidelines 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。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