The server has 3 explicitly defined tools. Two tools (get_detection_model_list, get_detection_model_details) have minimal or no input schemas and lack meaningful parameter descriptions. The nql_guide tool is read-only reference material. All three lack meaningful output documentation. Parameter naming and descriptions are sparse, and error handling is absent. The server relies heavily on OpenAPI auto-generation for many tools, which provides better structure than hand-written tools, but the three custom tools demonstrate poor definition quality.

• Add explicit descriptions to the model_name parameter in get_detection_model_details: 'The exact name of the detection model to retrieve. Call get_detection_model_list to see available options.'
• Document return types for all three tools. E.g., 'Returns: dict with keys: name (str), description (str), version (str), detection_types (list[str]).'
• Add error recovery guidance to get_detection_model_details: if a model is not found, the tool should return available models as suggestions rather than a bare 404.
• Expand nql_guide description to indicate when to call it: 'Comprehensive guide to NQL syntax, query structure, and best practices. Refer to this whenever constructing search queries or debugging failed queries.'
• Consider adding a 'readOnlyHint' annotation (per current MCP spec) to clarify that all three tools are safe read-only operations.
• For OpenAPI-generated tools, ensure customize_components() adds more specific descriptions to clarify search endpoint behavior (e.g., size parameter guidance is helpful, but extend descriptions of other auto-generated tools similarly).

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry