A collection of Skills, Commands, and Agents to help you get more out of Claude Code.
This repository provides three types of Claude Code resources:
Add-ons that teach Claude new capabilities. Each skill is a folder containing instructions and resources that Claude can use when needed.
Quick shortcuts that start with / to run common tasks. Type /command-name to instantly trigger a specific workflow.
Specialized helpers with focused expertise. Claude automatically brings in the right agent when it matches what you're trying to do.
This repository is organized by resource type:
claude-code-resources/
βββ README.md
βββ LICENSE
βββ skills/ # Claude Code Skills
β βββ skill-builder/
β βββ skill-contributor/
β βββ frontend-ui/
β βββ ui-ux-audit/
β βββ code-refactoring/
β βββ code-reviewer/
β βββ shownotes-generator/
β βββ feature-orchestrator/
βββ commands/ # Slash Commands
β βββ improve-code.md
β βββ plan-feature.md
β βββ README.md
βββ agents/ # Custom Subagents
βββ prompt-engineer.md
βββ README.md
Status: β
Available
Directory: skills/skill-builder/
Purpose: A meta-skill that helps you create well-structured Claude Code Skills
Features:
- Step-by-step skill creation workflow
- Templates for common patterns
- Validation tools
- Best practices guidance
- Progressive disclosure strategies
Quick Install:
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
cp -r claude-code-skills/skills/skill-builder ~/.claude/skills/Status: β
Available
Directory: skills/skill-contributor/
Purpose: Guides contributors through adding new skills to the repository via Pull Request
Features:
- Step-by-step PR submission workflow
- Automated validation checks
- README.md documentation updates
- Git branch and commit automation
- Pull Request template generation
- Post-submission checklist
Quick Install:
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
cp -r claude-code-skills/skills/skill-contributor ~/.claude/skills/Status: β
Available
Directory: skills/frontend-ui/
Purpose: Create aesthetically pleasing, visually distinctive frontend UIs using research-backed prompting strategies
Features:
- Four design dimensions (Typography, Color, Motion, Backgrounds)
- Avoids generic "AI slop" patterns
- Research-backed from Anthropic's frontend aesthetics cookbook
- 6+ theme aesthetics (Cyberpunk, Terminal, Brutalist, Solarpunk, etc.)
- Font rotation strategy for variety
- Comprehensive reference library
Quick Install:
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
cp -r claude-code-skills/skills/frontend-ui ~/.claude/skills/Status: β
Available
Directory: skills/ui-ux-audit/
Purpose: Enforces "READ FIRST, IMPLEMENT SECOND" workflow to prevent redundant UI/UX implementations
Features:
- Auto-invokes on UI/UX-related keywords (improve UI, enhance page, etc.)
- 6-step systematic audit workflow
- 7 pre-filled audit templates (quick, comprehensive, WCAG, etc.)
- UX research and WCAG 2.1 AA reference materials
- Real-world examples and case studies
- Prevents duplicate features and cluttered designs
- Integration with CLAUDE.md standards
Quick Install:
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
cp -r claude-code-skills/skills/ui-ux-audit ~/.claude/skills/Status: β
Available (v2.0.0)
Directory: skills/code-refactoring/
Purpose: AI-powered code complexity prevention with real-time monitoring and path-based intelligence
Features:
- π― NEW in v2.0: Path-based thresholds - Context-aware limits based on file purpose
- Pages: 300/500/800 lines (educational content allowed)
- Data files: 250/400/600 lines (static content)
- Components: 150/200/300 lines (standard threshold)
- Utilities: 100/150/200 lines (strict - should be small)
- π Real-time background file watcher with desktop notifications
- β¨ Automated execution - Step-by-step refactoring with your approval
- π Safety features - Incremental commits, automatic testing, rollback on failure
- π Multi-language support - JavaScript/TypeScript/React, Python, general patterns
- π₯οΈ Cross-platform - Windows, macOS, Linux with native scripts
- π 4 slash commands - Quick access to monitoring and scanning
- π Comprehensive resources - Code smells catalog, decision flowcharts, TDD integration
Auto-invokes when:
- Editing files >200 lines
- Creating files reaching 150 lines
- Reading files >300 lines
- Complexity patterns detected (4+ hooks, large classes, nested logic)
Quick Install:
# Clone repository
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
# Install skill (global - works across all projects)
cp -r claude-code-skills/skills/code-refactoring ~/.claude/skills/
# Install Node.js dependencies (required for file watcher)
cd ~/.claude/skills/code-refactoring/scripts
npm install
# (Optional) Install slash commands to project
cp ~/.claude/skills/code-refactoring/slashes-commands/*.md /path/to/project/.claude/commands/Status: β
Available
Directory: skills/shownotes-generator/
Purpose: Generate comprehensive "Shownotes" summaries from longform articles, papers, or transcripts of video/audio content
Features:
- Structured shownotes format with metadata, hooks, takeaways, and quotes
- Pauses to request metadata confirmation from user before full shownotes are generated
- Author/speaker attribution inferred from provided text
- Engaging hooks and key questions for target audiences
- Memorable quotes extraction and formatting
- Examples and reference materials included
Quick Install:
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
cp -r claude-code-skills/skills/shownotes-generator ~/.claude/skills/Status: β
Available
Directory: skills/feature-orchestrator/
Purpose: Research-backed feature implementation workflow enforcing gap analysis, incremental planning, agent coordination, and continuous integration best practices
Features:
- Phase 1: Gap Analysis - Search for existing code BEFORE writing anything new (prevents duplication)
- Phase 2: Implementation Planning - Break features into small, testable increments (<100 LOC each)
- Phase 3: Review Gate - Automatic critic-agent and ui-ux-designer invocation based on complexity
- Phase 4: Incremental Execution - "Never break the build" workflow with automated testing
- Research-backed by Google, Microsoft, IEEE, and Scrum.org (2024)
- File size monitoring with automated refactoring suggestions
- 3 automation scripts (validate-increment.sh/bat, check-file-size.sh)
- 3 templates (gap analysis, implementation plan, increment checklist)
- Comprehensive REFERENCE.md with research citations and detailed procedures
- Real-world EXAMPLES.md showing complete workflows
Auto-invokes when:
- User requests "implement X feature"
- User asks to "add Y functionality"
- User wants to "build Z component"
- Any new feature implementation request
Integration with CLAUDE.md:
- Enforces all mandatory rules (search first, plan before implementing, test between increments)
- Coordinates with code-refactoring, ui-ux-audit, devops-deployment, and qa-testing skills
Quick Install:
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
cp -r claude-code-skills/skills/feature-orchestrator ~/.claude/skills/Status: β
Available
Directory: skills/code-reviewer/
Purpose: Research-backed code review with OWASP Top 10 security checks, SAST tool integration, performance pattern detection, and automated quality standards enforcement
Features:
- OWASP Top 10 2021 Security Checks - Complete checklist for injection, XSS, authentication, access control, and more
- SAST Tool Integration - Works with SonarQube, CodeQL, Snyk, ESLint, npm audit
- Performance Pattern Detection - Identifies N+1 queries, O(nΒ²) algorithms, memory leaks, bundle bloat
- Code Quality Standards - TypeScript strict mode, naming conventions, complexity limits
- Severity Classification - Critical/High/Medium/Low with clear action items
- Automated Scripts - Cross-platform quick-audit scripts (bash + Windows batch)
- Research-backed - Based on OWASP 2025, CWE Top 25, empirical studies (92% faster remediation)
- Complete examples - Good vs bad review walkthroughs, security-focused reviews (EXAMPLES.md)
- Ready-to-use templates - 4 review templates in resources/templates/ (comprehensive, security, performance, quick)
- Real-world examples - Good/bad review examples in resources/examples/
- Complete reference - OWASP checklist, CWE patterns, language-specific guidance (REFERENCE.md)
Auto-invokes when:
- User mentions "review this code"
- User asks to "check for bugs"
- User requests "security audit"
- User wants to "analyze this PR"
- Any code review or quality check request
Manual invocation:
- Before committing critical changes
- Pre-deployment validation
- After implementing security-sensitive features
- When integrating SAST tool results
Quick Install:
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
cp -r claude-code-skills/skills/code-reviewer ~/.claude/skills/Purpose: Create a safe, step-by-step plan with explicit non-goals and risks
Creates comprehensive feature plans with goals, risks, atomic tasks, and auto-generates IMPLEMENTATION.md and USER_SETUP.md files.
Usage: /plan-feature user-authentication
Purpose: Smart code improvements with safety and performance focus
Performs targeted code improvements across components, APIs, database, or performance with safety flags and impact assessment.
Usage: /improve-code component --safe
Model: opus (most capable) Purpose: Create, modify, review, and optimize system prompts for AI agents
Expert prompt engineer that analyzes, designs, and enhances system prompts with effectiveness scoring, clarity analysis, and specific improvements.
Example: "Review this customer service agent prompt and make it more effective"
To install everything in this collection:
# Clone the repository
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
cd claude-code-skills
# Install all skills
cp -r skills/* ~/.claude/skills/
# Install all commands (when available)
cp commands/*.md ~/.claude/commands/ 2>/dev/null || true
# Install all agents (when available)
cp agents/*.md ~/.claude/agents/ 2>/dev/null || trueTo install specific items:
# Clone the repository
git clone https://github.com/WomenDefiningAI/claude-code-skills.git
# Install a specific skill
cp -r claude-code-skills/skills/skill-builder ~/.claude/skills/
# Install a specific command (when available)
cp claude-code-skills/commands/example-command.md ~/.claude/commands/
# Install a specific agent (when available)
cp claude-code-skills/agents/example-agent.md ~/.claude/agents/To install resources for a specific project only:
# From within your project directory
cp -r /path/to/claude-code-skills/skills/skill-builder .claude/skills/
cp /path/to/claude-code-skills/commands/example-command.md .claude/commands/Once installed, Claude will automatically discover and use skills when appropriate:
- "Use the skill-builder to create a new skill"
- "Help me build a code review skill"
- "Create a beautiful landing page" (uses frontend-ui skill)
Invoke commands with the slash prefix:
/plan-feature user-authentication
/improve-code component --safe
Claude automatically uses agents when they're needed. You don't have to do anything - Claude will bring them in when appropriate.
You can install resources in two places:
Use this for resources you want available everywhere. They'll work in all your projects.
Use this for resources specific to one project. They'll only work in that project folder.
- Claude Code CLI installed
- Python 3.8 or newer (only for some skills)
- Check individual skill READMEs for any other requirements
We welcome contributions from everyone! Whether you're a developer or new to open source, you can contribute skills, commands, and agents to this collection.
The easiest way to contribute is using our skill-contributor skill, which automates the entire process:
-
Install the skill-contributor:
git clone https://github.com/WomenDefiningAI/claude-code-skills.git cp -r claude-code-skills/skills/skill-contributor ~/.claude/skills/ -
Let Claude handle everything:
- Simply tell Claude: "Use the skill-contributor to add my new skill"
- The skill-contributor will guide you through creating a pull request
- No need to understand Git or GitHub - Claude does it for you!
Important: All contributions must be submitted via Pull Request. Do not commit directly to the main branch.
New to Git or GitHub? No problem!
- Share your idea as an Issue, or
- Use the skill-contributor (see above) - Claude handles the technical stuff for you, or
- Ask questions in Discussions - we're happy to help!
Manual steps:
- Fork this repository on GitHub
- Create a new branch for your contribution (never commit to main)
- Add your resource to the appropriate directory (
skills/,commands/, oragents/) - Update documentation (main
README.mdand relevant section READMEs) - Validate your skill structure (if contributing a skill)
- Submit a Pull Request from your branch
- Wait for review - maintainers may request changes
Before submitting your Pull Request:
- Created resource follows the documented format (YAML frontmatter, etc.)
- Included README.md with installation and usage instructions
- Updated main README.md to list your contribution
- Validated skill structure (if contributing a skill)
- Used a descriptive Pull Request title and description
- Submitted from a feature branch (not main/master)
We'd love to see what you create! Check out the Contributing section above to get started.
This collection is maintained by Women Defining AI to help developers work more effectively with Claude Code.
- Issues: Report bugs or request features via GitHub Issues
- Discussions: Ask questions or share ideas in GitHub Discussions
MIT License - see individual skill directories for specific licensing information.
Built with reference to:
Happy skill building! π