MCP server for Exa AI search API integration
Single tool 'search' has critical definition gaps. Input schema exists but parameters lack descriptions (both null). No output schema documented. No error handling guidance. No pagination despite API returning potentially large result sets. Fails baseline requirements: 100% of A+ tools have docstrings; 100% of A+ tool params have descriptions.
All input parameters lack descriptions (query and fullText both have null descriptions). LLMs cannot infer parameter meaning from names alone.
No output schema documented. LLMs cannot plan downstream calls or extract structured data from raw API response text.
No pagination support (limit, offset, page_size parameters missing). API response is returned as raw text without result count or next_cursor, risking context window exhaustion.
No error handling guidance. If API fails or returns invalid results, LLM receives raw error text with no recovery path (e.g., 'try a different query', 'check API key').
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 34 | 2026-07-28+ | v2 |