building-skills
Expert at creating and modifying Claude Code skills. Auto-invokes when the user wants to create, update, modify, enhance, validate, or standardize skills, or when modifying skill YAML frontmatter fields (especially 'allowed-tools', 'description'), needs help designing skill architecture, or wants to understand when to use skills vs agents. Also auto-invokes proactively when Claude is about to write skill files (*/skills/*/SKILL.md), create skill directory structures, or implement tasks that involve creating skill components.
SKILL.md
| Name | building-skills |
| Description | Expert at creating and modifying Claude Code skills. Auto-invokes when the user wants to create, update, modify, enhance, validate, or standardize skills, or when modifying skill YAML frontmatter fields (especially 'allowed-tools', 'description'), needs help designing skill architecture, or wants to understand when to use skills vs agents. Also auto-invokes proactively when Claude is about to write skill files (*/skills/*/SKILL.md), create skill directory structures, or implement tasks that involve creating skill components. |
Claude Code Plugins
A collection of plugins and meta-agents for Claude Code, Anthropic's official CLI for Claude. These plugins extend Claude's capabilities with specialized agents, skills, commands, and hooks.
CI/CD Pipeline: Automated validation, changelog management, and semantic versioning. See CI/CD Guide β
Now available on Claude Marketplace! Install with one command or browse plugins directly from Claude Code.
π What is Claude Code?
Claude Code is an interactive CLI tool that helps with software engineering tasks. It supports:
- Agents: Specialized subagents for delegated tasks
- Skills: Auto-invoked expertise modules
- Slash Commands: User-triggered workflows
- Hooks: Event-driven automation
- Plugins: Bundled collections of the above
π¦ Available Plugins
Agent Builder (Meta-Agent)
A meta-agent for building other Claude agents!
The Agent Builder plugin provides comprehensive tools for creating Claude Code extensions. It includes expert agents, auto-invoked skills, validation scripts, and templates for all component types.
Features:
- π€ Meta-architect agent for expert guidance
- π οΈ 4 specialized skills that auto-invoke when building components
- β‘ 5 slash commands for quick scaffolding
- β Validation scripts for schema compliance
- π Complete template library
- π Security-focused best practices
Quick Start:
# Create a new agent
/claude-component-builder:agents:new code-reviewer
# Create a new skill
/claude-component-builder:skills:new analyzing-data
# Create a new command
/claude-component-builder:commands:new run-tests
# Create a new hook
/claude-component-builder:hooks:new validate-writes
# Create a complete plugin
/claude-component-builder:plugins:new my-tools
Self-Improvement Plugin
Claude critiquing Claude - A feedback loop for continuous improvement!
The Self-Improvement plugin enables Claude to critique its own work, identify quality issues, and create feedback loops for continuous learning. It includes a self-critic agent and skills for quality analysis, improvement suggestions, and feedback loops.
Features:
- π Self-critic agent for honest, constructive feedback
- π Quality analysis across 6 dimensions (correctness, completeness, clarity, efficiency, security, usability)
- π‘ Actionable improvement suggestions with before/after examples
- π Feedback loop systems for continuous learning
- β‘ Quick quality checks and comprehensive reviews
- π Pattern tracking and learning from mistakes
Quick Start:
# Comprehensive review of recent work
/review-my-work
# Quick quality assessment
/quality-check
Use Cases:
- Catch bugs before delivery
- Learn from recurring mistakes
- Iterative refinement of solutions
- Security hardening through self-review
- Communication improvement
π₯ Installation
From Claude Marketplace (Recommended)
Install directly from the Claude Code marketplace:
# Add the marketplace to your Claude Code settings
claude marketplace add https://github.com/C0ntr0lledCha0s/claude-code-plugin-automations
# Install specific plugins
claude plugin install claude-component-builder
claude plugin install self-improvement
Or add to your .claude/settings.json:
{
"marketplaces": [
"https://github.com/C0ntr0lledCha0s/claude-code-plugin-automations"
],
"plugins": [
"claude-component-builder",
"self-improvement"
]
}
For All Projects (User-Level)
Install plugins to your user-level Claude directory:
# Clone the repository
git clone https://github.com/C0ntr0lledCha0s/claude-code-plugin-automations.git
cd claude-code-plugin-automations
# Link plugins to Claude user directory
mkdir -p ~/.claude/plugins
ln -s $(pwd)/claude-component-builder ~/.claude/plugins/claude-component-builder
ln -s $(pwd)/self-improvement-plugin ~/.claude/plugins/self-improvement
For a Specific Project
Install plugins at the project level:
# Navigate to your project
cd /path/to/your/project
# Clone the plugins
git clone https://github.com/C0ntr0lledCha0s/claude-code-plugin-automations.git
# Link components to project .claude directory
mkdir -p .claude/{agents,skills,commands}
ln -s $(pwd)/claude-code-plugin-automations/claude-component-builder/agents/* .claude/agents/
ln -s $(pwd)/claude-code-plugin-automations/claude-component-builder/skills/* .claude/skills/
ln -s $(pwd)/claude-code-plugin-automations/claude-component-builder/commands/* .claude/commands/
ln -s $(pwd)/claude-code-plugin-automations/self-improvement-plugin/agents/* .claude/agents/
ln -s $(pwd)/claude-code-plugin-automations/self-improvement-plugin/skills/* .claude/skills/
ln -s $(pwd)/claude-code-plugin-automations/self-improvement-plugin/commands/* .claude/commands/
Using plugin.json
If your Claude Code version supports it, use the plugin manifest:
# Add to your project's .claude/settings.json
{
"plugins": [
"/path/to/claude-code-plugin-automations/claude-component-builder",
"/path/to/claude-code-plugin-automations/self-improvement-plugin"
]
}
π― Use Cases
Building Custom Agents
Create specialized agents for your workflow:
- Code review agents
- Testing agents
- Documentation agents
- Security audit agents
- Refactoring agents
Creating Auto-Invoked Skills
Build skills that automatically activate:
- Data analysis expertise
- Language-specific knowledge
- Framework-specific guidance
- Domain-specific best practices
Automating Workflows
Design slash commands for common tasks:
- Git workflows (commit, push, PR creation)
- Testing and building
- Code generation and scaffolding
- Deployment and releases
Policy Enforcement
Implement hooks for validation:
- Prevent writes to protected directories
- Validate bash commands for security
- Auto-format code after changes
- Enforce coding standards
π Documentation
Quick Links
Claude Code Resources
ποΈ Project Structure
claude-code-plugin-automations/
βββ .claude-plugin/
β βββ marketplace.json # Marketplace manifest
βββ claude-component-builder/ # Meta-agent plugin
β βββ .claude-plugin/
β β βββ plugin.json # Plugin manifest
β βββ agents/ # Meta-architect agent
β βββ skills/ # 4 builder skills
β βββ commands/ # 5 creation commands
β βββ README.md
βββ self-improvement-plugin/ # Self-improvement plugin
β βββ .claude-plugin/
β β βββ plugin.json # Plugin manifest
β βββ agents/ # Self-critic agent
β βββ skills/ # Quality analysis skills
β βββ commands/ # Review commands
β βββ hooks/ # Feedback hooks
β βββ README.md
βββ README.md # This file
βββ LICENSE # MIT License
βββ .gitignore # Git ignore rules
ποΈ Marketplace Compatibility
This repository is fully compatible with Claude Code marketplaces! The .claude-plugin/marketplace.json file enables:
- Easy Discovery: Browse plugins from Claude Code
- One-Command Install: Install plugins with
claude plugin install <name> - Auto-Updates: Keep plugins up-to-date automatically
- Metadata: Rich plugin information including categories, keywords, and versions
- Strict Mode: Full validation against plugin manifests
Marketplace Metadata
- Total Plugins: 2
- Categories: Development Tools, Productivity
- All plugins include: Complete manifests, documentation, and examples
- Compliance: Follows 2025 Claude Code plugin standards
Contributing Improvements Back
The self-improvement plugin creates a meta-feedback loop where it can identify and contribute improvements to itself! See the complete workflow:
π Marketplace Contribution Workflow - Detailed guide on how improvements flow from pattern detection β implementation β marketplace distribution
Key Features:
- π Self-discovery of plugin limitations through pattern analysis
- π οΈ Structured improvement and testing workflow
- π¦ Version management and semantic versioning
- π Automated distribution through marketplace
- π Impact tracking and continuous measurement
This creates a true continuous improvement cycle where the plugins improve themselves based on usage patterns!
π§ Development
CI/CD Pipeline
This repository includes a comprehensive CI/CD pipeline for quality assurance:
Automated Validation:
- Runs on all PRs and commits to main/develop
- Validates all plugin components (agents, skills, commands, hooks)
- Enforces conventional commit format
- Checks JSON schema compliance
Changelog Management:
- Automatic changelog generation from commits
- Semantic versioning support
- GitHub release automation
Commands:
# Validate all plugins
npm run validate
# Quick validation
npm run validate:quick
# Generate changelog
npm run changelog
# Bump version (patch/minor/major)
npm run version:patch
npm run version:minor
npm run version:major
Learn more: CI/CD Guide
Creating Your Own Plugin
Use the Agent Builder to create new plugins:
/claude-component-builder:plugins:new my-awesome-plugin
Claude will guide you through:
- Defining the plugin structure
- Creating component directories
- Writing the manifest file
- Adding documentation
Contributing New Plugins
To add a plugin to this repository:
- Create your plugin using the Agent Builder
- Validate all components using the validation scripts
- Document your plugin with a comprehensive README
- Test all components thoroughly
- Submit a PR with your plugin in a new directory
Plugin Guidelines
All plugins should:
- Follow Claude Code naming conventions (lowercase-hyphens)
- Include a plugin.json manifest
- Provide comprehensive documentation
- Include validation scripts where applicable
- Follow security best practices
- Include usage examples
π§ͺ Testing
Validate Components
The Agent Builder includes validation scripts:
# Validate an agent
python claude-component-builder/skills/building-agents/scripts/validate-agent.py .claude/agents/my-agent.md
# Validate a skill
python claude-component-builder/skills/building-skills/scripts/validate-skill.py .claude/skills/my-skill/
# Validate a command
python claude-component-builder/skills/building-commands/scripts/validate-command.py .claude/commands/my-command.md
# Validate hooks
python claude-component-builder/skills/building-hooks/scripts/validate-hooks.py .claude/hooks.json
Manual Testing
Test components by:
- Creating test cases for agents, skills, and commands
- Triggering hooks with various scenarios
- Verifying auto-invocation of skills
- Checking error handling and edge cases
π€ Contributing
Contributions are welcome! Here's how you can help:
Report Issues
Found a bug or have a feature request? Open an issue
Submit Plugins
Have a useful plugin to share? Submit a PR with:
- Complete plugin directory
- Comprehensive README
- Validation scripts (if applicable)
- Usage examples
- Tests
Improve Documentation
Help improve:
- Plugin documentation
- Usage examples
- Best practices guides
- Tutorial content
Code Review
Review PRs from other contributors:
- Check for security issues
- Verify naming conventions
- Test functionality
- Suggest improvements
π Naming Conventions
All components follow strict naming conventions:
- Lowercase letters, numbers, and hyphens only
- No underscores or special characters
- Maximum 64 characters
Specific Conventions
- Plugins:
plugin-name(descriptive, unique) - Agents:
agent-name(action-oriented:code-reviewer,test-runner) - Skills:
skill-name(gerund form:analyzing-data,generating-reports) - Commands:
command-name(verb-first:run-tests,create-component) - Hooks: Event-based names (descriptive of validation/action)
π Security
Best Practices
When creating plugins:
- Minimize tool permissions: Start with Read, Grep, Glob
- Validate all inputs: Check for injection attacks
- Avoid hardcoded secrets: Use environment variables
- Review bash commands: Audit for dangerous operations
- Test security: Try to bypass your own validations
Reporting Security Issues
Found a security vulnerability? Please email security@example.com (or open a private security advisory)
π License
This project is licensed under the MIT License - see the LICENSE file for details.
π Acknowledgments
- Anthropic for creating Claude and Claude Code
- The Claude Code community for inspiration and feedback
- Contributors who help improve these plugins
π Links
π¬ Support
- Issues: GitHub Issues
- Discussions: GitHub Discussions
- Discord: Join our community (if applicable)
πΊοΈ Roadmap
Completed Plugins
- β Agent Builder Plugin: Meta-agent for building Claude Code extensions
- β Self-Improvement Plugin: Continuous improvement through self-critique and feedback loops
Future Plugin Ideas
- Testing Suite Plugin: Automated test generation and execution
- Documentation Plugin: Auto-generate docs from code
- Security Scanner Plugin: Vulnerability detection and reporting
- Performance Analyzer Plugin: Code optimization suggestions
- Git Workflow Plugin: Advanced git operations and automation
- API Generator Plugin: REST/GraphQL API scaffolding
- Database Plugin: Schema management and migration tools
- Code Review Plugin: Automated code review with best practices
- Refactoring Plugin: Intelligent code refactoring suggestions
Built with β€οΈ using Claude Code - A meta-agent building meta-agents! π€
Star History
If you find these plugins useful, please consider starring the repository! β