Add MiniMax as cloud LLM provider

[octo-patch]

Summary

Add MiniMax as a first-class cloud LLM provider, following the existing external model provider pattern alongside OpenAI, Ollama, Foundry Local, and Lemonade.

MiniMax offers OpenAI-compatible API endpoints at api.minimax.io/v1, enabling seamless integration using the existing Microsoft.Extensions.AI.OpenAI NuGet package (no additional dependencies needed).

Models included

• MiniMax-M2.7 - Latest flagship model with 1M context window
• MiniMax-M2.5 - 204K context window
• MiniMax-M2.5-highspeed - High-speed variant with 204K context

Changes (18 files, 790 additions)

Core provider:

• MiniMaxModelProvider.cs: Provider implementing IExternalModelProvider, uses OpenAI SDK with custom endpoint
• HardwareAccelerator.MINIMAX: Enum value added in all 3 locations (main, source generator, project template)
• ExternalModelHelper.cs: Registered MiniMax in provider list
• ModelDetailsHelper.cs: Added to IsLanguageModel() check

UI:

• MiniMaxPickerView.xaml + .xaml.cs: API key entry and model selection view (mirrors OpenAI picker pattern)
• ModelPickerDefinitions.cs: Registered MiniMax picker tab
• Icon assets: SVG (dark/light themes) + PNG variants

Tests:

• 10 unit tests: singleton, properties, NuGet refs, model listing, client creation, code snippet generation
• 3 integration tests: API connectivity, chat completion, streaming (gated by MINIMAX_API_KEY env var)

Docs:

• AddingSamples.md: Added MINIMAX to HardwareAccelerator values list

Test plan

• Build solution on Windows with Visual Studio 2022
• Run unit tests (MiniMaxModelProviderTests)
• Verify MiniMax tab appears in model picker alongside OpenAI
• Enter MiniMax API key and verify model list loads
• Select a MiniMax model and run a language model sample
• Run integration tests with MINIMAX_API_KEY set (optional)