MCP server for searching and retrieving patent information from Google Patents public dataset via BigQuery
Three tools with basic verb-noun naming (search_patents, get_patent, get_patent_claims). All have descriptions (50-100 chars, below the 194-char baseline). Input schemas visible in testclient but NOT in actual tool definitions, only inferred from test calls. No parameter descriptions in schemas. No output schema documented. No error handling guidance. No tool annotations. Composition is reasonable (three distinct operations), but schema completeness and parameter documentation are severely lacking.
Fetch full details of a patent by its publication number
Fetch the full claims text of a US patent by publication number. Claims are the legal scope of the patent.
Search patents by keyword, inventor, or technology area using Google Patents public dataset
Tool definitions not visible in source code. Tools inferred from testclient calls only; actual tool registration in tools/ package not provided. Cannot verify input schemas, parameter descriptions, or output structure.
Parameter descriptions missing. 'query' and 'publication_number' parameters have descriptions in testclient, but no evidence these are present in actual tool schema definitions. LLMs cannot infer parameter meaning from names alone.
No output schema documented. Tools return TextContent but no structured field definitions provided. LLMs cannot plan downstream calls or extract specific data from responses.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | <=2025-11-25 | v2 |
No error handling guidance. BigQuery client returns errors but no recovery hints provided. LLMs receive raw errors with no actionable next steps.
No tool annotations. Tools lack readOnlyHint, destructiveHint, or idempotentHint. All three are read-only but this is not declared in the tool definition.