[FEATURE] Support llmman as an OpenAI-compatible provider

Feature Area: Integration with external tools

Problem

llmman runs local models distributed as OCI artifacts and serves an OpenAI-compatible API on port 17434. CrewAI already ships first-class support for the other local OpenAI-compatible servers (Ollama, vLLM), but there is no way to point a crew at llmman without hand-configuring base_url on every LLM.

Proposal

Register llmman alongside the existing entries:

• A ProviderConfig in OPENAI_COMPATIBLE_PROVIDERS (LLMMAN_HOST, no API key required).
• Runtime routing in crewai.llm.LLM so LLM(model="llmman/qwen3.8") and LLM(model="qwen3.8", provider="llmman") reach OpenAICompatibleCompletion.
• CLI scaffolding entries (PROVIDERS, ENV_VARS, MODELS) so crewai create offers it.

LLMMAN_HOST is a bare host:port value, so it needs the same /v1 base-URL normalization already applied to OLLAMA_HOST.

Additional context

Happy to implement this.

Start with OPENAI_COMPATIBLE_PROVIDERS and the existing OLLAMA_HOST normalization, then trace provider routing in crewai.llm.LLM. Review the CLI PROVIDERS, ENV_VARS, and MODELS entries. Done means both model/provider forms route to OpenAICompatibleCompletion and crewai create exposes llmman with LLMMAN_HOST.