feat: add MiniMax as first-class LLM provider

[octo-patch]

Summary

Adds MiniMax as a built-in LLM provider for Tribe, enabling users to select MiniMax-M2.7 and MiniMax-M2.7-highspeed models directly from the provider dropdown.

Changes

Backend (backend/app/core/graph/members.py):

• Added minimax provider case in BaseNode.__init__ using ChatOpenAI with MiniMax's OpenAI-compatible API endpoint (https://api.minimax.io/v1)
• Auto-detects MINIMAX_API_KEY from environment variables
• Clamps temperature to (0, 1] range (MiniMax API constraint)
• Streaming enabled by default

Frontend (frontend/src/components/Members/EditMember.tsx):

• Added minimax entry to AVAILABLE_MODELS with MiniMax-M2.7 and MiniMax-M2.7-highspeed models
• Shows the "Proxy Provider" (base URL) field for MiniMax with default URL hint

Configuration & Docs:

• .env.example: Added MINIMAX_API_KEY= placeholder
• README.md: Added "Using MiniMax Models" section with setup instructions

Tests (backend/app/tests/graph/test_minimax_provider.py):

• 13 unit tests covering provider initialization, temperature clamping, model names, API key handling, and GraphMember/GraphTeam compatibility
• 3 integration tests (skipped when MINIMAX_API_KEY is not set)

Test Plan

• Unit tests verify ChatOpenAI instance creation with correct base_url, model, temperature, and API key
• Temperature clamping tested for edge cases: 0.0→0.01, 1.0→1.0, 1.5→1.0, -0.5→0.01
• Integration test verified MiniMax API responds correctly via OpenAI-compatible endpoint
• Manual: select minimax provider in the Edit Member dialog, verify model dropdown shows M2.7 models
• Manual: verify base URL field appears with default hint for MiniMax provider