Implement Model Context Protocol (MCP) integration with Aspire orchestration

[Copilot]

This PR implements a comprehensive Model Context Protocol (MCP) solution integrated with .NET Aspire, enabling AI models to dynamically discover and use tools through a standardized interface.

What is MCP?

Model Context Protocol (MCP) is a standard that enables:

• Dynamic Tool Discovery: AI models can discover available tools at runtime
• Standardized Interface: Consistent API for tool calling across different systems
• Interoperable Hosting: Tools can be reused across multiple AI applications

Key Features Implemented

🔧 MCP Server (BuildWithAspire.McpServer)

• Weather Tool: Get forecasts for any location with customizable days (1-7)
• Time Tool: Current date/time with timezone support and multiple formats
• REST API: Standardized endpoints for tool discovery (/mcp/tools) and execution (/mcp/tools/call)
• JSON Schema Validation: Comprehensive parameter validation and error handling

🤖 AI Service Integration

• Intelligent Message Parsing: Automatically detects weather/time queries in chat
• MCP Client Service: Seamless communication with MCP server
• Context Enrichment: AI responses enhanced with real tool data
• Demo Endpoint: Live demonstration at /mcp/demo

🏗️ Aspire Orchestration

• Service Discovery: AI service automatically locates MCP server
• Health Monitoring: Built-in health checks and telemetry
• Scalable Architecture: Independent scaling of tool servers

✅ Comprehensive Testing

• 16 Unit Tests: 100% passing test suite covering all functionality
• Parameter Validation: Tests for JSON deserialization and edge cases
• Error Scenarios: Comprehensive error handling validation

Usage Examples

Intelligent Chat Integration

# Ask for weather - automatically calls MCP weather tool
POST /conversations/{id}/messages
{"message": "What's the weather like in Tokyo?"}

# Ask for time - automatically calls MCP time tool  
POST /conversations/{id}/messages
{"message": "What time is it in New York?"}

Direct Tool Calls

# Call weather tool directly
POST /mcp/tools/call
{
  "name": "get_weather",
  "arguments": {
    "location": "San Francisco",
    "days": 3
  }
}

# Get available tools
GET /mcp/tools

Benefits Over Traditional APIs

Traditional APIs	MCP Integration
❌ Hardcoded endpoints	✅ Dynamic discovery
❌ Manual integration	✅ Standardized interface
❌ Tight coupling	✅ Loose coupling via protocol
❌ Limited reusability	✅ Cross-system compatibility

Architecture

┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│   AI Service    │────│   MCP Server    │────│   Tool Suite    │
│ - Chat Service  │    │ - Tool Registry │    │ - Weather Tool  │
│ - AI Models     │    │ - Request/Resp  │    │ - Time Tool     │
│ - MCP Client    │    │ - Validation    │    │ - Custom Tools  │
└─────────────────┘    └─────────────────┘    └─────────────────┘
         │                       │                       │
         └───────────────────────┼───────────────────────┘
                                 │
                    ┌─────────────────┐
                    │ Aspire AppHost  │
                    │ - Service Disc. │
                    │ - Health Checks │
                    │ - Telemetry     │
                    └─────────────────┘

Documentation

Added comprehensive documentation (MCP_INTEGRATION.md) covering:

• Detailed architecture explanation
• Tool development patterns
• Best practices and security considerations
• Migration guide from traditional APIs
• Future enhancement roadmap

Migration Path

The implementation maintains backward compatibility:

• ✅ Original /weatherforecast endpoint still works
• ✅ Existing chat functionality preserved
• ✅ Progressive adoption of MCP tools

This establishes a foundation for building sophisticated AI applications with rich, discoverable tool ecosystems using industry-standard protocols.

Fixes #10.

Warning

Firewall rules blocked me from connecting to one or more addresses

I tried to connect to the following addresses, but was blocked by firewall rules:

• https://api.github.com/search/repositories
  • Triggering command: curl -s REDACTED (http block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to my firewall allow list

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.