Skills Overview
48 specialized skills that extend Claude’s capabilities—loaded dynamically when you mention them.
Quick Reference
Frontend & Design
| Skill | Purpose |
|---|---|
| frontend-design | Build memorable web interfaces with bold aesthetics |
| ui-ux-pro-max | Production-ready UI with research-backed patterns |
| ui-styling | Tailwind patterns, responsive layouts, dark mode |
| frontend-development | React patterns, Suspense, state management |
| web-frameworks | Next.js + Turborepo + RemixIcon stack |
| threejs | 3D web experiences with WebGL/WebGPU |
| react-best-practices | React and Next.js performance optimization from Vercel |
| web-design-guidelines | Web Interface Guidelines compliance |
| web-testing | Playwright, Vitest, k6 testing |
| shader | GLSL fragment shaders for procedural graphics |
| remotion | Video creation in React |
Backend & Infrastructure
| Skill | Purpose |
|---|---|
| backend-development | Node.js, NestJS, security, testing patterns |
| databases | Schema design, query optimization, migrations |
| devops | CI/CD, deployment, infrastructure automation |
AI & Multimodal
| Skill | Purpose |
|---|---|
| ai-multimodal | Gemini vision, audio, document processing |
| google-adk-python | Google AI Development Kit for Python agents |
| ai-artist | Generate images via Nano Banana with curated prompts |
Tools & Utilities
| Skill | Purpose |
|---|---|
| mcp-builder | Create MCP servers (Python FastMCP / TypeScript) |
| mcp-management | Discover and execute MCP tools |
| skill-creator | Create custom skills for your projects |
| repomix | Pack repos into AI-friendly context files |
| document-skills | PDF, DOCX, PPTX, XLSX processing |
| docs-seeker | Find and retrieve external documentation |
| chrome-devtools | Browser automation, performance profiling |
| media-processing | FFmpeg, ImageMagick, background removal |
| agent-browser | AI-optimized browser automation |
| markdown-novel-viewer | Book-like markdown reading experience |
| context-engineering | Monitor and optimize token consumption |
| gkg | GitLab Knowledge Graph semantic analysis |
Process & Methodology
| Skill | Purpose |
|---|---|
| planning | Transform requirements into executable plans |
| research | Multi-source validation before implementation |
| sequential-thinking | Numbered thought sequences for complex problems |
| problem-solving | Systematic approaches when stuck |
| debug | Root cause investigation framework |
| code-review | Verification gates and technical rigor |
| brainstorm | Trade-off analysis with brutal honesty |
| scout | Fast parallel codebase scouting |
| cook | Feature implementation workflow |
| fix | Bug fixing with intelligent routing |
| git | Git operations with conventional commits |
| plans-kanban | Visual plan progress tracking |
Integrations & Specialized
| Skill | Purpose |
|---|---|
| better-auth | TypeScript auth (OAuth, 2FA, passkeys, multi-tenant) |
| shopify | Shopify apps, GraphQL API, checkout extensions |
| payment-integration | Stripe, PayPal, LemonSqueezy, SePay integration |
| mobile-development | React Native, Flutter, Swift/SwiftUI, Kotlin |
| copywriting | Conversion copy formulas and templates |
| mermaidjs-v11 | Diagram creation with Mermaid.js |
| find-skills | Discover and install agent skills |
How to Use
Basic invocation:
"Use [skill-name] to [task]"Examples:
"Use better-auth to add GitHub OAuth with 2FA"
"Use devops to create production deployment"
"Use debug to investigate this test failure"
"Use frontend-design to build a SaaS landing page"Skill not activating? Be explicit:
"Use the [skill-name] skill to..."Under the Hood
How Skills Activate
Skills activate through semantic matching on your prompt:
- Claude matches your request to skill descriptions
- Relevant skill instructions load into context
- Claude follows skill-specific patterns and best practices
Activation triggers:
- Mentioning the skill name explicitly
- Describing a task that matches skill description
- Using keywords from skill’s domain
Skill Structure
Every skill contains:
.claude/skills/[skill-name]/
├── SKILL.md # Core instructions (<100 lines)
├── references/ # Detailed documentation
└── scripts/ # Automation scripts (optional)Progressive disclosure: SKILL.md provides essentials, references/ has depth.
Skills vs Commands vs Agents
| Aspect | Skills | Commands | Agents |
|---|---|---|---|
| Purpose | Specialized knowledge | Workflow orchestration | Task execution |
| Invocation | ”Use [skill]…” | /command | Auto or explicit |
| Scope | Single focused capability | Multi-step processes | Autonomous work |
| Example | better-auth, devops | /plan, /test | planner, tester |
Creating Custom Skills
"Use skill-creator to create a skill for [your-domain]"skill-creator will:
- Ask clarifying questions
- Design skill structure
- Create SKILL.md with proper frontmatter
- Add references if needed
- Save to
.claude/skills/
Troubleshooting
Skill not working?
- Check skill name spelling
- Provide more context about your task
- Try explicit invocation: “Use the X skill to…”
Need a skill that doesn’t exist?
- Use skill-creator to build it
- Request on Discord
Key Takeaway
50 skills provide instant expertise—just mention the skill and describe your task. No configuration needed.