Communitygithub.com

modbender/skill-library-mcp

Use this skill when the user asks to build, create, design, develop, or improve ANY frontend interface, web page, UI component, or visual element. This includes: - Building landing pages, websites, web apps, dashboards, portfolios, or any web interface - Creating UI components (buttons, forms, cards, navbars, modals, etc.) - Designing pages with React, Vue, Next.js, Svelte, or any frontend framework - Adding styling or improving visual design of existing components - Implementing specific design aesthetics (modern, dark, minimalist, brutalist, etc.) - User mentions "frontend", "UI", "UX", "design", "interface", "web design", or "styling" - User asks for "beautiful", "modern", "professional", "clean", or any aesthetic adjective - User requests help with CSS, Tailwind, styled-components, or any styling approach This skill automatically retrieves the appropriate design system prompt (Neo-brutalism, Modern Dark, Bauhaus, Cyberpunk, Material, etc.) to help create visually distinctive, production-grade frontend...

What is skill-library-mcp?

skill-library-mcp is a Claude Code agent skill that use this skill when the user asks to build, create, design, develop, or improve ANY frontend interface, web page, UI component, or visual element. This includes: - Building landing pages, websites, web apps, dashboards, portfolios, or any web interface - Creating UI components (buttons, forms, cards, navbars, modals, etc.) - Designing pages with React, Vue, Next.js, Svelte, or any frontend framework - Adding styling or improving visual design of existing components - Implementing specific design aesthetics (modern, dark, minimalist, brutalist, etc.) - User mentions "frontend", "UI", "UX", "design", "interface", "web design", or "styling" - User asks for "beautiful", "modern", "professional", "clean", or any aesthetic adjective - User requests help with CSS, Tailwind, styled-components, or any styling approach This skill automatically retrieves the appropriate design system prompt (Neo-brutalism, Modern Dark, Bauhaus, Cyberpunk, Material, etc.) to help create visually distinctive, production-grade frontend...

Works with✓Claude Code~Codex CLI~Cursor
npx skills add https://github.com/modbender/skill-library-mcp/tree/HEAD/data/design-style

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

Design Style Skill

Purpose

This skill helps Claude Code create beautiful, distinctive frontend interfaces by automatically retrieving design system prompts from the prompts/ directory. Instead of producing generic UI, this skill enables Claude to build interfaces with specific design aesthetics like Neo-brutalism, Modern Dark, Luxury, Cyberpunk, and more.

When to Use

This skill is automatically invoked when:

  • User asks to build a web page, landing page, or web application
  • User requests a UI component with a specific design style
  • User mentions frontend, React, Vue, or web development
  • User asks for a specific aesthetic (e.g., "make it look modern and dark" or "use a brutalist style")

Available Design Styles

The following design systems are available in the prompts/ directory:

  • Academia - Scholarly, classic, refined
  • ArtDeco - Luxurious 1920s glamour
  • Bauhaus - Functionalist, geometric minimalism
  • BoldTypography - Type-driven design
  • Botanical - Nature-inspired, organic
  • Claymorphism - Soft, clay-like 3D elements
  • Cyberpunk - Futuristic, neon, high-tech
  • Enterprise - Professional, corporate, scalable
  • FlatDesign - Clean, minimal, 2D
  • Fluent2 - Microsoft Fluent 2 Design System
  • HumanistLiterary - Warm, literary, conversational (Claude aesthetic)
  • Industrial - Raw, mechanical, utilitarian
  • Kinetic - Dynamic, motion-focused
  • Luxury - Premium, elegant, sophisticated
  • Material - Google Material Design
  • Maximalism - Bold, expressive, abundant
  • MinimalDrak - Minimal dark theme (note: typo in original)
  • ModernDark - Contemporary dark UI with depth
  • Monochrome - Black and white, high contrast
  • Neo-brutalism - Bold, raw, colorful brutalism
  • Neumorphism - Soft UI, skeuomorphic
  • Newsprint - Newspaper-inspired
  • Organic - Natural, flowing forms
  • PlayfulGeometric - Fun geometric shapes
  • Professional - Clean, business-focused
  • Retro - Vintage, nostalgic
  • Saas - Modern SaaS aesthetic
  • Sketch - Hand-drawn, artistic
  • Swiss - International Typographic Style
  • TerminalCLI - Command-line interface aesthetic
  • Vaporwave - 80s/90s aesthetic, nostalgic
  • Web3 - Decentralized, crypto-inspired

How It Works

Step 1: Understand User Intent

When the user requests frontend work, first determine:

  1. Tech stack - What framework are they using? (React, Vue, Next.js, etc.)
  2. Design preference - Did they mention a specific style or aesthetic?
  3. Component scope - Single component, full page, or entire application?

Step 2: Select Design Style

If user specifies a style:

  • Match their request to available styles (e.g., "brutalist" → Neo-brutalism)
  • Case-insensitive matching (brutalism, Brutalism, BRUTALISM all work)

If user doesn't specify:

  • For modern, professional projects → ModernDark or Professional
  • For creative, bold projects → Neo-brutalism or BoldTypography
  • For minimal, clean projects → FlatDesign or Swiss
  • For enterprise/corporate → Enterprise

Ask the user if you're uncertain about which style fits their needs.

Step 3: Retrieve Design System

Use the Read tool to load the appropriate prompt file:

Read: prompts/<StyleName>.md

For example:

  • prompts/Neo-brutalism.md
  • prompts/ModernDark.md
  • prompts/Cyberpunk.md

Step 4: Apply Design System

Once you've loaded the design system prompt:

  1. Internalize the design philosophy - Understand the core principles, visual signatures, and differentiation factors
  2. Extract design tokens - Colors, typography, spacing, shadows, borders
  3. Follow component patterns - Use the specified button styles, card layouts, etc.
  4. Apply the "Bold Factor" - Implement signature elements that make the design authentic
  5. Avoid anti-patterns - Don't use techniques that break the aesthetic

Step 5: Build with Context

Before writing code:

  • Identify the user's existing tech stack
  • Understand their component architecture
  • Note any constraints (CSS frameworks, design libraries, etc.)

When writing code:

  • Match their existing patterns and conventions
  • Centralize design tokens in CSS variables or a config file
  • Create reusable, composable components
  • Explain your architectural choices briefly

Quality standards:

  • Preserve or improve accessibility
  • Ensure responsive design across devices
  • Make deliberate, creative design choices (not generic boilerplate)
  • Leave the codebase cleaner than you found it

Examples

Example 1: User Specifies Style

User: "Create a landing page for my SaaS product with a neo-brutalist design"

Skill Actions:

  1. Detect keywords: "landing page", "neo-brutalist"
  2. Map "neo-brutalist" → prompts/Neo-brutalism.md
  3. Read the design system prompt
  4. Ask clarifying questions: "What tech stack are you using? React, Vue, or plain HTML/CSS?"
  5. Build the landing page following Neo-brutalism principles (thick borders, hard shadows, bold colors, etc.)

Example 2: User Doesn't Specify Style

User: "Help me build a portfolio website"

Skill Actions:

  1. Detect: "portfolio website" (creative context)
  2. Suggest options: "Would you like a specific design style? I can create it in Modern Dark (sophisticated), Neo-brutalism (bold and creative), or Swiss (minimal and clean)."
  3. User responds with preference
  4. Load appropriate prompt and build

Example 3: Component Request

User: "Add a contact form to my Next.js app. Make it look modern and professional."

Skill Actions:

  1. Keywords: "Next.js", "modern and professional"
  2. Select: ModernDark.md (modern) or Professional.md (professional)
  3. Read design system
  4. Build form component matching their Next.js patterns
  5. Use design tokens from the prompt (colors, typography, shadows, etc.)

Quick Reference Commands

When implementing, you can quickly reference specific sections:

Colors:

Grep: pattern "Token|Value|Usage" path "prompts/<Style>.md"

Typography:

Grep: pattern "Font|Weight|Size" path "prompts/<Style>.md"

Component Patterns:

Grep: pattern "Button|Card|Input" path "prompts/<Style>.md"

Tips for Best Results

  1. Read the full prompt - Don't just skim. The design philosophy and differentiation sections are critical.
  2. Implement signature elements - Every design system has a "Bold Factor" section. These elements are what make it authentic.
  3. Avoid anti-patterns - Each prompt lists what NOT to do. Follow these rules strictly.
  4. Ask questions - If the user's needs are unclear, ask before building.
  5. Match existing code - Don't fight their tech stack. Integrate the design system into their existing patterns.

Notes

  • All design system prompts follow the same structure: <role> and <design-system> sections
  • Prompts include detailed color palettes, typography scales, component patterns, and implementation examples
  • Each style is production-ready and has been carefully crafted for visual distinctiveness
  • The prompts are designed to prevent generic, AI-looking interfaces

Future Enhancements

Potential improvements to this skill:

  • Style combination support (e.g., "Cyberpunk + Minimal")
  • Custom style creation workflow
  • Design token extraction to JSON/CSS
  • Component library generation from prompts

Individual skills in this repo

This repo contains 20 individual skills — each has its own dedicated page.

modbender/skill-library-mcp

Paid acquisition strategy, budget allocation, and avoiding common advertising mistakes across platforms

modbender/skill-library-mcp

Complete investment portfolio risk management system. Analyze positions, calculate risk metrics, stress test scenarios, optimize allocations, and generate institutional-grade risk reports — all wit...

modbender/skill-library-mcp

Borrow from Aave via credit delegation. Agent self-funds by borrowing against delegator collateral. Supports borrow, repay, health checks. Works on Aave V2/V3.

modbender/skill-library-mcp

Discover and register autonomous AI agents. Use when: browsing the agent registry, submitting an agent for verification, or embedding agent badges. NOT for: general web search or agent capability analysis.

modbender/skill-library-mcp

Create AI avatar and talking head videos with OmniHuman, Fabric, PixVerse via inference.sh CLI. Models: OmniHuman 1.5, OmniHuman 1.0, Fabric 1.0, PixVerse Lipsync. Capabilities: audio-driven avatars, lipsync videos, talking head generation, virtual presenters. Use for: AI presenters, explainer videos, virtual influencers, dubbing, marketing videos. Triggers: ai avatar, talking head, lipsync, avatar video, virtual presenter, ai spokesperson, audio driven video, heygen alternative, synthesia alternative, talking avatar, lip sync, video avatar, ai presenter, digital human

modbender/skill-library-mcp

AI代码编辑器规则模板集合 - 为Cursor、Windsurf、Claude Code、Cline等AI编辑器提供项目规则配置。适用于需要配置AI编码助手规则的开发者,包含全栈Web、移动端、Vue3+SpringBoot等技术栈模板。

modbender/skill-library-mcp

Generate landing page code from product descriptions

modbender/skill-library-mcp

Create AI marketing videos for ads, promos, product launches, and brand content. Models: Veo, Seedance, Wan, FLUX for visuals, Kokoro for voiceover. Types: product demos, testimonials, explainers, social ads, brand videos. Use for: Facebook ads, YouTube ads, product launches, brand awareness. Triggers: marketing video, ad video, promo video, commercial, brand video, product video, explainer video, ad creative, video ad, facebook ad video, youtube ad, instagram ad, tiktok ad, promotional video, launch video

modbender/skill-library-mcp

Generate AI music videos end-to-end. Creates music with Suno (sunoapi.org), generates visuals with OpenAI/Seedream/Google/Seedance, and assembles into music video with ffmpeg. Supports timestamped lyrics (auto SRT), Suno native music video generation, slideshow/video/hybrid modes. Token-based cost tracking per generation.

modbender/skill-library-mcp

Generate AI-powered notes from videos (document, outline, or graphic-text formats)

modbender/skill-library-mcp

The video AI notes tool is provided by Baidu. Based on the video download address provided by the user, it downloads and parses the video, and finally generates AI notes corresponding to the video (a total of three types of notes can be generated: document notes, outline notes, and image-text notes).

modbender/skill-library-mcp

Automated appointment management for beauty salons, clinics, studios, and photo booths. Handles booking requests, calendar sync, conflict detection, reminders, no-show tracking, and waitlist management.

modbender/skill-library-mcp

Bird identification, life list tracking, and trading card generation. Use this skill when the user: sends a bird photo to identify, says "set up my Birdfolio" or "set my region", asks "how's my checklist", asks "Birdfolio progress", asks "how many birds have I found", asks "show my Birdfolio" or "show my life list", asks "what's my rarest bird", or asks "tell me about [bird species]". Handles everything from first-time setup through ongoing life list tracking and visual trading card generation.

modbender/skill-library-mcp

Multi-project task and research management (JSON-first CLI)

modbender/skill-library-mcp

Track stock and crypto portfolios with CLI - real-time prices, P&L, dividend tracking, multiple portfolios

modbender/skill-library-mcp

Track cryptocurrency portfolios with real-time prices, P&L calculations, and allocation analysis. Query Bitcoin, Ethereum, Solana and 10,000+ token holdings.

modbender/skill-library-mcp

加密貨幣Portfolio追蹤 - 支援TRON/ETH/BSC,分析持倉、收益、Gas費用

modbender/skill-library-mcp

Lead strategy with competitive analysis, strategic planning, growth roadmaps, and business portfolio management.

modbender/skill-library-mcp

Create ATS-optimized resumes and CVs with proper structure, metrics, and industry targeting.

modbender/skill-library-mcp

Native video analysis using Google Gemini API. Upload and analyze video files — describe scenes, extract text/UI, answer questions about content, transcribe speech, identify objects and actions. Use when: (1) User sends a video file and wants it analyzed, (2) Video summarization or description needed, (3) Extracting text, UI elements, or information from screen recordings, (4) Answering questions about video content, (5) Comparing multiple videos, (6) Analyzing tutorials, demos, or walkthroughs.

Related Skills