Development Guide
Welcome to the CAIPE development documentation! This section provides comprehensive guides for developers who want to contribute to the project or build their own agents and MCP servers.
📚 What You'll Learn
This development guide covers:
- Spec-Driven Development - How CAIPE uses spec-kit for agentic software development
- Development Environment Setup - Configure your local development environment
- Creating an Agent - Build a new AI agent from scratch
- Creating an MCP Server - Develop a Model Context Protocol server
- CI/CD and Releases - Understand PR version bumps, prebuild artifacts, RC tags, hotfixes, and final releases
🎯 Who Is This For?
These guides are designed for:
- Contributors wanting to add new features or fix bugs
- Developers building code-defined agents for their organization
- Platform Engineers extending the platform with new capabilities
- Students learning about AI agents and MCP
🛠️ Prerequisites
Before you begin, you should have:
- Basic knowledge of Python 3.14
- Familiarity with async/await patterns
- Understanding of REST APIs
- Git and Docker installed
- An LLM API key (OpenAI, Anthropic, or Azure OpenAI)
🚀 Quick Start Path
- Understand spec-driven development - How CAIPE uses spec-kit for feature development
- Set up your development environment - Install dependencies and configure your workspace
- Explore Agent Builder - Create a governed agent without writing application code
- Create a code-defined agent - Build an agent when you need full application control
- Add an MCP server - Connect your agent to external APIs
📖 Additional Resources
- Contributing Guide - Contribution workflow and standards
- Architecture Overview - System design and components
- Agents & MCP Servers - Existing implementations
- Security Best Practices - Security guidelines
💡 Development Philosophy
CAIPE follows these principles:
- Convention over Configuration - Sensible defaults with flexibility
- Security by Default - Built-in authentication and authorization
- Observable & Traceable - OpenTelemetry integration out of the box
- Test-Driven - Comprehensive testing with evaluations
- Production-Ready - Docker, Kubernetes, and CI/CD support
🏅 Contribution Badges
CAIPE recognizes your journey with four badge levels — from running your first multi-agent setup to deploying at production scale and mentoring others.
| Badge | Level | How to Earn |
|---|---|---|
| 🧑🚀 Explorer | 1 | Run the multi-agent system with at least one use case |
| 🔧 Improver | 2 | Customize agents, add RAG, integrate tracing and evaluations |
| 🧠 Expert | 3 | Zero-trust deployment, agent gateways, OSS contributions |
| 🥋 Master | 4 | Production-scale deployment, mentorship, pioneering new patterns |
See the full badge guide and how to claim yours → Contributing & Badges
🤝 Getting Help
- Issues: GitHub Issues
- Discussions: GitHub Discussions
- Community: CAIPE community meetings
- Slack: Join the CAIPE community channel on CNCF Slack
Let's build something amazing together! 🚀