feat: add MiniMax provider support

[octo-patch]

Summary

Add MiniMax as a first-class LLM provider in the multi-provider system, supporting MiniMax-M2.5 and MiniMax-M2.5-highspeed models via the OpenAI-compatible API.

Changes

• New file: v3/@claude-flow/providers/src/minimax-provider.ts — Full MiniMax provider implementation extending BaseProvider
• Modified: types.ts — Add 'minimax' to LLMProvider union type and MiniMax models to LLMModel
• Modified: provider-manager.ts — Register MiniMax in the provider factory switch
• Modified: index.ts — Export MiniMaxProvider with updated module docs
• Modified: provider-integration.test.ts — Integration tests for completion, streaming, and model listing
• Modified: README.md — Add MiniMax to provider lists, architecture diagram, and environment variables

MiniMax Provider Details

• API: OpenAI-compatible (https://api.minimax.io/v1)
• Models: MiniMax-M2.5 (default) and MiniMax-M2.5-highspeed
• Context: 204,800 tokens, up to 192K output
• Features: Chat completion, streaming, tool calling
• Temperature: Clamped to (0.0, 1.0] range (MiniMax rejects zero)
• Auth: MINIMAX_API_KEY environment variable

Pricing (per 1M tokens)

Model	Input	Output
MiniMax-M2.5	$0.30	$1.20
MiniMax-M2.5-highspeed	$0.60	$2.40

API Documentation

• OpenAI Compatible: https://platform.minimax.io/docs/api-reference/text-openai-api

Testing

• All 3 MiniMax integration tests pass (completion, streaming, model listing)
• Build passes with no TypeScript errors

[ruvnet]

Thanks for the MiniMax provider PR @octo-patch. This looks well-structured with proper tests.

We're currently focused on bug fixes and stability for v3.5.x. Provider additions will be reviewed in the next feature cycle. Keeping this open for future consideration.

Note: #1324 (duplicate MiniMax PR by @ximiximi423) has been closed in favor of this one.

[octo-patch]

Thanks @ruvnet! Totally understand the focus on stability for v3.5.x. Happy to keep this open and address any feedback when the feature cycle comes around.

And thanks for consolidating with #1324 — makes sense to keep things clean.