devops-ai-first-phase by BaruchiHalamish20 - MCP Server

🤖 MCP Agent System - Python Implementation

A practical implementation of Model Context Protocol (MCP) with Agent-to-Agent (A2A) communication using Python and Ollama. Perfect for learning and teaching AI agent architectures.

📖 Overview

This project demonstrates:

MCP Server: Provides reusable tools (weather, time, calculator)
Coordinator Agent: Plans and delegates tasks
Worker Agent: Executes tasks using MCP tools
A2A Communication: Agents collaborate to solve problems

Architecture

┌─────────────────┐
│  User Request   │
└────────┬────────┘
         │
         ▼
┌─────────────────────┐
│ Coordinator Agent   │  ◄── Plans & Decides
│   (Ollama LLM)      │
└─────────┬───────────┘
          │ A2A Communication
          ▼
┌─────────────────────┐
│   Worker Agent      │  ◄── Executes Tasks
│   (Ollama LLM)      │
└─────────┬───────────┘
          │
          ▼
┌─────────────────────┐
│    MCP Server       │  ◄── Tool Execution
│  (weather, time,    │
│   calculator)       │
└─────────────────────┘

🚀 Quick Start

Prerequisites

Python 3.10 or higher
Ollama installed and running
Basic understanding of async Python

Installation

Clone or create the project:

mkdir mcp-agents-demo
cd mcp-agents-demo

Set up Python environment:

python3 -m venv venv
source venv/bin/activate  # On Windows: venv/Scripts/activate

Install dependencies:

pip install mcp httpx

Create the files:

Copy mcp_server.py (MCP server implementation)
Copy agents.py (Agent system)
Copy requirements.txt

Install and run Ollama:

# Install Ollama (if not already installed)
# curl -fsSL https://ollama.com/install.sh | sh

# Start Ollama server
ollama serve

# In another terminal, pull a model
# ollama pull llama2  # or llama3, mistral, etc.

Running the Demo

python3 agents.py

You'll see:

✅ Coordinator analyzing requests
✅ Agent-to-agent communication
✅ MCP tool execution
✅ Complete results

📂 Project Structure

mcp-agents-demo/
├── mcp_server.py       # MCP server with tools
├── agents.py           # Two-agent system
├── requirements.txt    # Python dependencies
└── README.md          # This file

🎯 How It Works

1. MCP Server (`mcp_server.py`)

Provides three tools accessible via MCP protocol:

Tool	Description	Arguments
`get_weather`	Get weather for a city	`city: string`
`get_time`	Get current time	`timezone: string`
`calculate`	Math operations	`operation, a, b`

2. Coordinator Agent

Role: High-level task planning and delegation

Responsibilities:
- Analyze user requests
- Understand available tools
- Create execution plans
- Delegate to Worker Agent

Example Flow:

User: "What's the weather in Tokyo?"
↓
Coordinator: Analyzes → Identifies "get_weather" tool needed
↓
Delegates to Worker with plan

3. Worker Agent

Role: Task execution using MCP tools

Responsibilities:
- Receive tasks from Coordinator
- Select appropriate MCP tools
- Execute tool calls
- Return results

Example Flow:

Receives: "Get weather for Tokyo"
↓
Calls: get_weather(city="Tokyo")
↓
Returns: Weather data

4. A2A Communication

Agents communicate by sharing:

Task context
Execution plans
Conversation history
Tool results

# Example A2A flow
coordinator_result = await coordinator.process(request)
↓
worker_result = await worker.execute_task(
    coordinator_result["task"],
    coordinator_result["decision"]  # A2A context sharing
)

🔧 Configuration

Change Ollama Model

Edit agents.py:

OLLAMA_MODEL = "llama3"  # Options: llama2, llama3, mistral, codellama

Customize Test Cases

Edit the test_requests list in agents.py:

test_requests = [
    "What is the weather in Tokyo?",
    "Calculate 15 divided by 3",
    "What time is it in Paris?",
    "Your custom request here"
]

Add New MCP Tools

In mcp_server.py, add to handle_list_tools():

types.Tool(
    name="your_tool",
    description="Tool description",
    inputSchema={
        "type": "object",
        "properties": {
            "param": {"type": "string"}
        },
        "required": ["param"]
    }
)

Then implement in handle_call_tool():

elif name == "your_tool":
    result = your_implementation(arguments)
    return [types.TextContent(type="text", text=json.dumps(result))]

📊 Example Output

======================================================================
🚀 Starting MCP Agent System (A2A Demo)
======================================================================

✅ Connected to MCP server

======================================================================
📝 TEST CASE 1: What is the weather in Tokyo?
======================================================================

🎯 COORDINATOR AGENT: Analyzing request...
   Request: What is the weather in Tokyo?

🤖 Calling Ollama...

📋 Coordinator Analysis:
   This request requires the get_weather tool...

----------------------------------------------------------------------

🔄 AGENT-TO-AGENT COMMUNICATION
   Coordinator → Worker: Delegating task

⚙️  WORKER AGENT: Executing task...
   Task: What is the weather in Tokyo?

📡 Calling MCP tool: get_weather(city='Tokyo')
✅ Tool Result: {
  "city": "Tokyo",
  "temperature": 23,
  "condition": "Sunny",
  "humidity": 65
}

======================================================================
✅ FINAL RESULT:
{
  "success": true,
  "result": "...",
  "tool": "get_weather"
}
======================================================================

🧪 Testing

Run Default Tests

python3 agents.py

Add Interactive Mode

Add to agents.py:

async def interactive_mode():
    mcp_client = MCPClientWrapper()
    await mcp_client.connect()
    
    coordinator = CoordinatorAgent(mcp_client)
    worker = WorkerAgent(mcp_client)
    
    print("Interactive mode. Type 'quit' to exit.\n")
    
    while True:
        user_input = input("Your request: ")
        if user_input.lower() in ['quit', 'exit']:
            break
        
        coord_result = await coordinator.process(user_input)
        worker_result = await worker.execute_task(
            coord_result["task"],
            coord_result["decision"]
        )
        
        print("\n✅ Result:")
        print(json.dumps(worker_result, indent=2))
        print()
    
    await mcp_client.close()

if __name__ == "__main__":
    asyncio.run(interactive_mode())

🐛 Troubleshooting

Ollama Connection Issues

Problem: Warning: Ollama not running

Solution:

# Start Ollama in a separate terminal
ollama serve

# Verify it's running
curl http://localhost:11434/api/tags

MCP Server Connection Failed

Problem: Cannot connect to MCP server

Solutions:

Ensure mcp_server.py is in the same directory
Check Python path permissions
Verify mcp package is installed: pip install mcp

Module Import Errors

Problem: ModuleNotFoundError: No module named 'mcp'

Solution:

# Activate virtual environment
source venv/bin/activate

# Install/reinstall dependencies
pip install --upgrade mcp httpx

Model Not Found

Problem: Ollama model not available

Solution:

# List available models
ollama list

# Pull the model you need
ollama pull llama2
# or
ollama pull llama3

🎓 Learning Exercises

Add a Database Tool
- Create query_database tool in MCP server
- Add SQL query execution capability
- Test with sample queries
Create a Validator Agent
- Add third agent to validate results
- Implement result checking logic
- Chain: Coordinator → Worker → Validator
Implement Error Recovery
- Add retry logic to Worker Agent
- Handle MCP tool failures gracefully
- Log errors for debugging
Add Persistent Memory
- Store conversation history
- Enable context across sessions
- Implement memory retrieval
Build Web Interface
- Create FastAPI/Flask endpoint
- Add REST API for agent system
- Build simple web UI

🔬 Advanced Features

Enable Debug Logging

export MCP_DEBUG=1
python3 agents.py

Run with Different Models

# Edit agents.py to change OLLAMA_MODEL
# Then run specific tests
python3 agents.py

Benchmark Agent Performance

Add timing to agents.py:

import time

start = time.time()
result = await coordinator.process(request)
end = time.time()

print(f"Coordinator took {end - start:.2f}s")

📚 Key Concepts

Model Context Protocol (MCP)

Standardized protocol for AI tool integration
Enables tool reusability across different agents
Separates tool implementation from agent logic

Agent-to-Agent Communication

Agents share context and decisions
Enables complex multi-agent workflows
Maintains conversation history

Tool Calling Pattern

Agent identifies needed tool
Agent prepares tool arguments
Tool executes via MCP
Result returned to agent
Agent processes and responds

🤝 Contributing

Ideas for contributions:

Add more MCP tools (file operations, API calls, etc.)
Improve LLM prompt engineering
Add unit tests
Create visualization dashboard
Improve error handling

📖 Resources

📝 License

This project is provided as-is for educational purposes. Feel free to use, modify, and distribute.

💬 Support

For questions or issues:

Check the troubleshooting section
Review MCP documentation
Test with simplified examples
Verify Ollama is running correctly

🎯 Next Steps

✅ Run the basic demo
✅ Understand agent communication flow
✅ Modify test cases
✅ Add custom tools
✅ Experiment with different models
✅ Build your own agent system!

Happy Learning! 🚀

Built for DevOps engineers working with AI agents and MCP.

BaruchiHalamish20/devops-ai-first-phase

🤖 MCP Agent System - Python Implementation

📖 Overview

Architecture

🚀 Quick Start

Prerequisites

Installation

Running the Demo

📂 Project Structure

🎯 How It Works

1. MCP Server (mcp_server.py)

2. Coordinator Agent

3. Worker Agent

4. A2A Communication

🔧 Configuration

Change Ollama Model

Customize Test Cases

Add New MCP Tools

📊 Example Output

🧪 Testing

Run Default Tests

Add Interactive Mode

🐛 Troubleshooting

Ollama Connection Issues

MCP Server Connection Failed

Module Import Errors

Model Not Found

🎓 Learning Exercises

🔬 Advanced Features

Enable Debug Logging

Run with Different Models

Benchmark Agent Performance

📚 Key Concepts

Model Context Protocol (MCP)

Agent-to-Agent Communication

Tool Calling Pattern

🤝 Contributing

📖 Resources

📝 License

💬 Support

🎯 Next Steps

1. MCP Server (`mcp_server.py`)