feat: add MiniMax as first-class LLM provider

[octo-patch]

Summary

This PR adds MiniMax as a first-class AI provider in Agentfy, following the same pattern as the existing ChatGPT and Claude wrappers.

Changes

• common/ais/minimax.py — MiniMax async client wrapping the OpenAI-compatible API (https://api.minimax.io/v1). Supports M2.7, M2.7-highspeed, M2.5, M2.5-highspeed (all 204K context). Includes temperature clamping to (0.0, 1.0] as required by the API, model cost tracking, and model-name normalisation.
• common/ais/__init__.py — exports the new MiniMax class alongside ChatGPT and Claude.
• common/exceptions/exceptions.py — adds MiniMaxAPIError following the existing exception hierarchy.
• config.py — adds minimax_api_key field reading from MINIMAX_API_KEY environment variable.
• tests/test_minimax_unit.py — 22 unit tests covering init, temperature clamping, model normalisation, cost calculation, and error handling (no network required).
• tests/test_minimax_integration.py — 3 integration tests (auto-skipped when MINIMAX_API_KEY is not set).
• tests/conftest.py — pytest path setup.

Usage

from common.ais.minimax import MiniMax

mm = MiniMax()  # reads MINIMAX_API_KEY from env
result = await mm.chat(
    system_prompt="You are a helpful assistant.",
    user_prompt="Summarise the latest social media trends.",
    model="MiniMax-M2.7",      # 204K context
)

MiniMax offers competitive pricing and a large 204K context window, making it an excellent addition to Agentfy's multi-provider AI toolkit.