Goal

Prepare and maintain the core NocoBase AI runtime and saved LLM services required by AI employees and, after the commercial capability gate has passed, knowledge bases. Require explicit direct-CLI-versus-UI routing for LLM service and vector database connection configuration, and keep chat-model configuration, embedding-model discovery, credentials, dependency checks, and verification separate and explicit.

Scope

• Confirm the target nb environment, authentication, runtime refresh, and generated ai command surface.
• Discover LLM providers, unsaved provider models, saved services, chat models, and embedding models.
• Test provider settings and create, update, enable, disable, verify, or safely delete saved LLM services.
• Before creating or changing an LLM service or vector database connection configuration, require the user to choose direct CLI parameters or the documented UI flow; verify user-completed UI work before follow-up.
• Produce a structured core-AI readiness result for nocobase-ai-employee.
• Produce conditional LLM/embedding prerequisites for nocobase-ai-knowledge-base-manager only after that skill has confirmed knowledge-base capability.
• Check employee and knowledge-base dependencies before disruptive LLM service changes.

Non-Goals