MCP server that analyzes API documentation HTML to extract endpoints, infer TypeScript schemas, detect authentication patterns, and test endpoints
Four tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. All tools have descriptions (10-20 chars baseline met), but parameter descriptions are sparse or missing. Input schemas use Zod but lack formal descriptions for most parameters. No output schemas documented. Error handling is absent, tools return raw JSON without recovery guidance. Naming is verb-forward (parse_, infer_, detect_, test_) but descriptions are terse and lack WHEN/WHY context. No parameter constraints (enums, ranges) despite tools accepting free-form strings (method, headers). The test_endpoint tool accepts arbitrary HTTP methods and bodies with no validation hints.
Detect the authentication pattern from HTTP response headers.
Infer a TypeScript type string from one or more JSON response examples.
Parse raw HTML from an API docs page and extract a structured list of endpoints.
Send a real HTTP request to verify an endpoint is live and see its response.
No output schemas documented. Tools return raw JSON strings without field definitions. LLMs cannot plan downstream calls or extract structured data reliably.
Parameter descriptions missing or incomplete. 'examples' in infer_schema has no description. 'method', 'url', 'headers', 'body' in test_endpoint lack constraints or format hints. LLMs cannot infer valid values.
No input validation or error guidance. test_endpoint accepts any HTTP method string with no enum constraint. No error responses explain what went wrong or how to recover.
Tool descriptions are terse (35-50 chars). Missing WHEN/WHY context. 'Send a real HTTP request to verify an endpoint is live' does not explain when to call this vs other tools or what the response structure contains.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 41 | 2026-07-28+ | v2 |
No pagination or result limits documented. parse_docs may return unbounded endpoint lists. No guidance on handling large responses or chunking results.