Skill Llm Minimax Cn

Skill LLM MiniMax China Mainland

Purpose

Define how an agent integrates a China Mainland MiniMax model already named by the user or host project. Keep this region separate because domains, credentials, availability, and pricing may differ.

Read Order

Read shared contracts first:

1. ../_shared/route-key-schema.md
2. ../_shared/model-catalog-schema.md
3. ../_shared/pricing-matrix-schema.md
4. ../_shared/capability-matrix-schema.md
5. ../_shared/connection-profile-schema.md
6. ../_shared/request-url-matrix-schema.md
7. ../_shared/request-envelope.md
8. ../_shared/response-envelope.md
9. ../_shared/error-contract.md
10. ../_shared/progress-contract.md
11. ../_shared/ui-binding.md
12. ../_shared/sync-policy.md
13. ../_shared/logging-fields.md

Then read:

1. references/api-surface-scope.md
2. references/connection-profiles.md
3. references/request-urls.md
4. references/model-catalog.md
5. references/pricing-matrix.md
6. references/capability-matrix.md
7. the transport for the requested kind
8. references/logging-contract.md when logging is implemented
9. references/model-sync.md only for an explicit update or verification task

Read references/build-multimodal-http.md only for endpoint references not represented by a shared request kind.

Runtime Rules

• Use provider identifier minimax-cn and the official base URL recorded in the resolved profile.
• Send the exact model ID from model-catalog.md.
• Resolve connection profile, request kind, API surface, API version, endpoint kind, full route key, base URL, and request URL before capability lookup.
• Match capabilities by the exact full RouteKey row, including API version and endpoint kind.
• chat and vision use OpenAI-compatible POST /v1/chat/completions with MiniMax-M3.
• imaging uses POST /v1/image_generation with image-01.
• music uses POST /v1/music_generation with music-2.6.
• MiniMax-M3 supports adaptive thinking, streaming, function tools, and image/video message input. This skill maps image input; video requires a separate typed host input contract.
• max_completion_tokens for MiniMax-M3 has a documented recommended value of 131072 and maximum of 524288.
• Treat build and plan as credential/profile names, not model policy. Enforce only the request-kind restrictions recorded in connection-profiles.md.
• Do not switch between https://api.minimaxi.com/v1 and the other MiniMax region.
• Do not infer Anthropic-compatible behavior from the OpenAI-compatible rows.
• Reject requested fields marked unsupported or unknown; do not silently drop them.
• For music streaming, stream=true requires hex output.
• Video endpoints remain reference-only until the shared request and response contracts define RequestKind = video.

Request Flow

1. Read the requested region, model, request kind, and profile.
2. Resolve the exact request URL row.
3. Resolve the exact catalog and capability rows.
4. Validate thinking, tools, media, stream, image, and music fields.
5. Apply the matching transport.
6. Normalize the provider response, errors, progress, and logs through the shared contracts.

Request Kinds

• chat: text-first Chat Completions
• vision: image understanding through M3 message content
• imaging: image generation through image-01
• music: music generation through music-2.6

Out of Scope

• the other MiniMax region
• Anthropic-compatible MiniMax surfaces
• first-class video integration before the shared schema defines it
• model ranking or permission policy
• non-MiniMax providers

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: joeylu
• Source: joeylu/runtime-free-llm-integration-skills
• License: MIT

Install and usage instructions live in the source repository linked above.