feat: add MiniMax as a first-class LLM provider

[octo-patch]

Summary

• Add MiniMax as a first-class LLM provider with dedicated adapter
• Temperature clamping ensures values stay within MiniMax's valid range [0, 1.0]
• Full streaming support via OpenAI-compatible API
• 11 unit tests + 3 integration tests covering all adapter functionality

Changes

File	Description
adapter/minimax/chat.go	ChatInstance with streaming, temperature clamping, error handling
adapter/minimax/struct.go	Request/Response type definitions
adapter/minimax/chat_test.go	11 unit tests
adapter/minimax/integration_test.go	3 integration tests (build-tagged)
globals/constant.go	MinimaxChannelType constant
adapter/adapter.go	Factory registration
README.md / README_zh-CN.md / README_ja-JP.md	Add MiniMax to supported models list

MiniMax API Details

• API Base: https://api.minimax.io/v1 (OpenAI-compatible)
• Models: MiniMax-M2.7, MiniMax-M2.5, MiniMax-M2.5-highspeed (up to 204K context)
• Temperature: Clamped to [0, 1.0] range automatically

Test Plan

• All 11 unit tests pass (go test ./adapter/minimax/ -v)
• All 3 integration tests pass with live MiniMax API (go test ./adapter/minimax/ -tags=integration)
• Project builds successfully (go build ./...)
• Configure MiniMax channel in admin dashboard with endpoint and API key