Specification of a completion handler function with asynchronous execution support.
Completions generate AI model outputs based on prompt or resource references and
user-provided arguments. This abstraction enables:
- Customizable response generation logic
- Parameter-driven template expansion
- Dynamic interaction with connected clients
requests and returns results. The first argument is an
McpAsyncServerExchange used to interact with the client. The second
argument is a
McpSchema.CompleteRequest.