Add C# ChatModelRouterMultiturn demo using Microsoft.Extensions.AI

[Copilot]

Ports the Python chat-model-router-multiturn.py demo to C# following the existing agent runner pattern in extensions-ai.

Changes

ChatModelRouterMultiturn.cs (new)

• Interactive multi-turn chat with model router showing which model handled each request
• History trimming: preserves system message + last N user/assistant pairs
• Uses Microsoft.Extensions.AI.ChatMessage alias to resolve OpenAI namespace conflict
• Configurable via appsettings.json with safe defaults

Program.cs

• Added option 3 for ChatModelRouterMultiturn
• Configuration parsing with TryParse and null-safe defaults

appsettings.json

• Added ModelRouter and chat parameters as native JSON types

Usage

// From Program.cs
var modelRouterMultiturn = new ChatModelRouterMultiturn();
await modelRouterMultiturn.Run(
    modelRouter,      // "model-router" deployment
    endpoint, 
    systemPrompt,
    maxHistoryTurns,  // default: 12
    maxOutputTokens,  // default: 10000
    temperature,      // default: 0.7
    topP,             // default: 0.95
    frequencyPenalty, 
    presencePenalty);

The implementation maintains functional parity with the Python version while using Microsoft.Extensions.AI abstractions consistent with other demos in the project.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>C# Model Router Multiturn</issue_title>
<issue_description>demos\01-essentials\01-foundry\02-sdks\foundry-sdk-py\chat-model-router-multiturn.py contains a python implementation. I want to tranform this into c# in demos\01-essentials\01-foundry\02-sdks\extensions-ai

Keep the agent runner pattern</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes C# Model Router Multiturn #82

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.