feat(providers): upgrade MiniMax provider to support M2.7 models

[octo-patch]

Summary

• Add MiniMax-M2.7 and MiniMax-M2.7-highspeed models alongside existing M2.5 models
• M2.7 features a 1M token context window (vs 204K for M2.5) and 131K max output tokens
• Fix temperature clamping to allow temperature=0 (now accepted by MiniMax API)
• Remove unnecessary default temperature=1.0 when not specified by caller

Changes

File	Change
minimax-provider.ts	Add M2.7/M2.7-highspeed to supported models, fix temp clamping
types.ts	Add M2.7 model type definitions
index.ts	Update module documentation
README.md	Update MiniMax API key description
minimax-provider.test.ts	24 new unit tests (capabilities, model info, temp, errors, responses)
provider-integration.test.ts	Add M2.7 integration tests (completion, streaming, model listing)

Model Specifications

Model	Context	Max Output	Speed
MiniMax-M2.7	1,048,576	131,072	Standard
MiniMax-M2.7-highspeed	1,048,576	131,072	Fast
MiniMax-M2.5	204,800	192,000	Standard
MiniMax-M2.5-highspeed	204,800	192,000	Fast

Test Plan

• 24 unit tests passing (capabilities, model info, temperature clamping, error handling, response transformation)
• 4 integration tests passing with real MiniMax API (M2.7 completion, M2.5 completion, M2.7 streaming, model listing)
• All existing tests unaffected

Test Results

✓ @claude-flow/providers/src/__tests__/minimax-provider.test.ts (24 tests) 294ms
✓ @claude-flow/providers/src/__tests__/provider-integration.test.ts (4 passed | 10 skipped) 11598ms