Turn OpenAPI and Swagger specs into validated, curated, safe MCP tools.
Single tool 'dynamic_tools_from_openapi' has critical definition gaps. Input schema is declared as empty object with additionalProperties=true and a generic description stating schemas are 'dynamically generated per endpoint', this is not a valid, inspectable schema. Tool description is adequate (89 chars) but lacks guidance on when/why to use it. No parameter descriptions visible. The tool name lacks a clear action verb (should be 'call_openapi_endpoint' or similar). Tool definitions are inferred from OpenAPI specs at runtime, not explicitly registered with proper schemas, triggering the 50-point cap for inferred tools. No output schema documented. No error handling guidance.
Dynamically generated tools from OpenAPI/Swagger specifications. Each tool corresponds to an API endpoint defined in the spec, with names derived from operationId or method+path, descriptions from summary/description fields, and input schemas from parameters and request bodies.
Input schema is empty object with additionalProperties=true; no inspectable parameter definitions.
Tool name 'dynamic_tools_from_openapi' lacks action verb. Should start with verb (e.g., 'call_', 'invoke_', 'execute_') to signal intent to LLM.
Tool definitions are dynamically generated from OpenAPI specs at runtime, not explicitly registered with static schemas.
No output schema documented. LLMs cannot predict response structure or plan downstream tool calls.
No error handling guidance. Tool description does not explain what to do if an endpoint fails, returns invalid data, or requires authentication.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 31 | 2026-07-28+ | v2 |