feat: add MiniMax as alternative LLM provider

[octo-patch]

Summary

• Add MiniMax M2.7 and M2.7-highspeed as alternative LLM providers via OpenAI-compatible API
• Full integration across Python CLI package (settings, config, CLI command) and TypeScript web frontend (provider template)
• 21 unit tests + 3 integration tests covering settings registry, config management, token verification, and live API calls

Changes

Python package (7 files, 141 additions)

File	Change
settings.py	MiniMaxM27Settings and MiniMaxM27HighspeedSettings presets with temperature=0.7 and openai_api_base
config.py	set_minimax_token() / load_minimax_token() with env var (MINIMAX_API_KEY) and config file support
cli.py	skyagi config minimax subcommand + runtime API key injection when MiniMax model is selected
util.py	verify_minimax_token() validation

TypeScript web (2 files)

File	Change
model.ts	MiniMax in ModelProvider enum + providerTemplates entry with M2.7 and M2.7-highspeed
.env.example	MINIMAX_API_KEY

Tests (3 files, 248 additions)

• 21 unit tests: settings registry, preset values, config load/save, env var precedence, token verification, factory pattern
• 3 integration tests: live M2.7 and M2.7-highspeed chat completions, settings-based model loading

How it works

MiniMax provides an OpenAI-compatible API, so the integration reuses LangChain ChatOpenAI class with a custom openai_api_base. No new LLM wrapper class is needed.

Test plan

• All 21 unit tests pass
• All 3 integration tests pass with live MiniMax API
• MiniMax models appear in skyagi model list output
• skyagi config minimax saves token to ~/.skyagi/config.json
• Manual test: run full simulation with MiniMax M2.7 model