Learning Path
Overview
The five exercises are designed to build your skills progressively, from basic agent patterns to advanced distributed systems concepts.
Prerequisites
Before starting, you should have basic knowledge of:
- Python or JavaScript (pick one language and stick with it)
- Command-line usage (cd, ls, git, pip/npm commands)
- APIs and HTTP (GET/POST requests)
- Git basics (clone, add, commit, push)
If you're not comfortable with any of these, spend 1-2 hours on foundational tutorials first.
The Five-Exercise Journey
🎯 Phase 1: Fundamentals (Exercises 1-2)
Exercise 1: React Foundation
Duration: 30-45 minutes
Difficulty: ⭐ Beginner
Learn the ReAct pattern: Reason → Act → Observe → Repeat
You'll build an agent that:
- Reasons about a problem
- Selects and calls tools
- Observes results
- Iterates to reach a goal
Skills gained:
- Agent architecture
- Tool pattern
- LLM integration
- Real-time feedback loops
Exercise 2: Jenkins MCP Server
Duration: 45-60 minutes
Difficulty: ⭐⭐ Intermediate
Create your first Model Context Protocol (MCP) server.
You'll build a server that:
- Exposes Jenkins capabilities to agents
- Implements standard MCP interface
- Handles tool calls from agents
- Manages build pipelines
Skills gained:
- MCP concepts
- Server architecture
- Jenkins API integration
- Tool definition
🎯 Phase 2: Integration (Exercises 3-4)
Exercise 3: Watcher Agent
Duration: 60-90 minutes
Difficulty: ⭐⭐ Intermediate
Build an intelligent agent that monitors pipelines.
You'll implement:
- Multi-server agent
- Polling and event handling
- Decision logic based on state
- Action triggers
Skills gained:
- Complex agent logic
- Multiple tool coordination
- State management
- Event-driven architecture
Exercise 4: Gitea MCP Server
Duration: 60-90 minutes
Difficulty: ⭐⭐⭐ Advanced
Create an MCP server for Gitea (Git platform).
You'll built:
- Git-aware MCP server
- Repository operations
- Webhook integration
- Complex tool chains
Skills gained:
- Advanced MCP patterns
- Version control integration
- Webhook handling
- Multi-operation workflows
🎯 Phase 3: Capstone (Exercise 5)
Exercise 5: Grand Finale
Duration: 90-180 minutes
Difficulty: ⭐⭐⭐⭐ Expert
Integrate everything into one unified system.
You'll create:
- Multi-server orchestration
- Event sourcing with SQS
- Complex workflows spanning all five services
- Production-grade error handling
Skills gained:
- System design
- Distributed systems
- Event sourcing
- Production automation
Recommended Timeline
| Week | Exercises | Daily Time | Total |
|---|---|---|---|
| 1 | Ex 1, Ex 2 | 1-2 hours | 4-5 hours |
| 2 | Ex 3, Ex 4 | 1.5-2 hours | 4-6 hours |
| 3 | Ex 5 | 2-3 hours | 4-6 hours |
Total time: 12-17 hours spread over 3 weeks
How to Use This Learning Path
Starting Fresh
- Start with Exercise 1
- Complete exercises in order (1 → 2 → 3 → 4 → 5)
- Don't skip exercises; each builds on previous knowledge
Coming Back?
- If you've done Exercise A before, you can skip to Exercise B+1
- But read the setup sections to refresh your memory
Need a Challenge?
- Try implementing advanced variants after each exercise
- See the "Advanced" section in each exercise README
- Build your own MCP servers
Stuck?
- Check the exercise README first
- See Troubleshooting
- Review example solutions
- Ask on GitHub or with your instructor
Learning Goals by Phase
After Exercise 1
✅ Understand agent reasoning and action loops
✅ Use tools to solve problems
✅ Integrate with LLMs
After Exercise 2
✅ Build MCP services
✅ Expose tools to agents
✅ Design tool interfaces
After Exercise 4
✅ Coordinate multiple services
✅ Handle complex workflows
✅ Manage state and events
After Exercise 5
✅ Build production systems
✅ Design for scale
✅ Implement event-driven architecture
✅ You're ready for real-world AI systems!
Next Step
Ready to begin? Head to Exercise 1: React Foundation →