feat: add MiniMax provider support

[octo-patch]

Summary

• Adds MiniMax as an OpenAI-compatible LLM provider (chat models only)
• Default model: MiniMax-M2.7, also supports MiniMax-M2.7-highspeed
• API endpoint: https://api.minimax.io/v1 (OpenAI-compatible protocol)
• Temperature default: 1.0 (MiniMax requires > 0.0)
• Context window: 1,000,000 tokens
• API key stored securely in Keychain under com.agent.minimax-api-key

Note: This PR depends on macOS26/AgentTools#1 which adds the APIProvider.miniMax enum case to the shared package.

Files changed

File	Change
Agent/Services/LLMProviderSetup.swift	Register MiniMax provider config
Agent/Services/KeychainService.swift	Add Keychain get/set for MiniMax API key
Agent/AgentViewModel/Core/AgentViewModel.swift	Add MiniMax key, model, models, temperature properties
Agent/AgentViewModel/Features/DefaultModels.swift	Add default MiniMax model list
Agent/AgentViewModel/Features/ModelFetching.swift	Add fetchMiniMaxModels() + fetchModelsIfNeeded case
Agent/AgentViewModel/Features/ScriptTabs.swift	Wire up globalModelForProvider, apiKeyForProvider, modelDisplayName
Agent/AgentViewModel/Core/Colors.swift	Add temperature binding
Agent/AgentViewModel/TaskExecution/Setup.swift	Add provider/model/vision resolution
Agent/Views/Settings/SettingsView.swift	Add MiniMax settings section (API key + model picker)
Agent/Views/Tabs/NewMainTabSheet.swift	Add MiniMax to new tab model picker
Agent/Views/Settings/FallbackChainView.swift	Add MiniMax to fallback chain default model
Agent/Views/Output/ThinkingIndicatorView.swift	Add 1M context window for MiniMax

Test plan

• Build succeeds with no Swift compiler errors (all exhaustive switch statements updated)
• MiniMax section appears in Settings under LLM Providers
• API key saves/loads correctly from Keychain
• Model picker shows MiniMax-M2.7 and MiniMax-M2.7-highspeed
• MiniMax appears as an option in New LLM Tab sheet
• MiniMax appears in Fallback Chain provider picker
• Sending a chat message to MiniMax returns a response