feat: add Crusoe model provider

[acheamponge]

Description

Adds Crusoe Managed Inference as a model provider, mirroring the existing Nebius provider layout.

Crusoe provides Managed Inference, an OpenAI-compatible API for open-weight models (GLM 5.2, DeepSeek V3, gpt-oss-120b, Gemma 4, Kimi K2.6, Llama 3.3) on Crusoe's vertically integrated AI cloud. Crusoe is already a supported provider in LiteLLM, Pydantic AI, Pipecat, aisuite, the Vercel AI SDK, and the models.dev registry.

Changes

• core/llm/llms/Crusoe.ts: provider class extending OpenAI with the Crusoe base URL (same shape as Nous.ts / Nebius.ts)
• core/llm/llms/index.ts: registration
• core/llm/autodetect.ts: added crusoe alongside nebius in the templating, images, and parallel-providers lists
• docs/customize/model-providers/more/crusoe.mdx + docs/docs.json: provider docs page and nav entry
• gui/.../configs/models.ts + providers.ts: Add New Model entries (GLM 5.2, Gemma 4 31B, and a Crusoe option on the existing gpt-oss-120b package)

Testing

The provider class is a direct mirror of Nous.ts/Nebius.ts, which pass through to the OpenAI implementation. The Crusoe endpoint itself has been verified OpenAI-compatible against the live API (chat, streaming, and tool-call requests, including for the models.dev entry). I have not run the Continue extension end-to-end against it, so happy to make any changes if something needs adjusting, and to run any additional verification you want.

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[acheamponge]

recheck

[acheamponge]

I have read the CLA Document and I hereby sign the CLA