golem-annotate-agent-scala

Import prompt and description from golem.runtime.annotations, then place @prompt and @description directly above each method in your agent trait. The @prompt tells an LLM when to call the method, while @description explains its behavior and parameters.

@prompt is a short natural-language instruction telling an LLM when to invoke the method, matched against user requests. @description is a longer explanation of what the method does, including parameters, return values, and edge cases.

No, @prompt and @description are purely metadata and have no effect on runtime behavior. They exist only to support AI and LLM tool discovery, so annotated and unannotated methods execute identically.

No, both annotations are optional and should be omitted for internal methods not intended for LLM discovery. Annotate only the methods you want external AI systems to find and invoke as tools.

Yes, agents with annotated methods can be used as tools by LLM-based systems. The annotations provide the discovery metadata an LLM needs to match user requests to the appropriate agent method.