An MCP server that integrates Claude AI with web search capabilities via DuckDuckGo, providing a tool to fetch web content based on user queries
Single tool 'fetch_web_content' has a basic schema and minimal description. Description is only 54 characters, below the 10 - 1024 character baseline and lacks WHEN/WHY context. Schema is present but parameter 'query' has no type constraint (string type is declared but no format, length, or enum guidance). No output schema documented. No error handling guidance. Tool does not follow verb_noun naming (should be 'search_web' or 'get_web_content'). No pagination, no result limits documented. Implementation is custom HTTP, not standard MCP protocol.
Retrieves info from website based on user queries
Tool description is only 54 characters and lacks actionable context. Should explain WHAT it does, WHEN to use it, and what it returns. Current: 'Retrieves info from website based on user queries', too vague.
No output schema documented. LLM cannot plan downstream calls or extract structured data. Must document return type (e.g., {content: string, url: string, status: string}).
Parameter 'query' lacks constraints: no min/max length, no format guidance, no enum. Description should state expected format (URL vs search term vs both?) and any length limits.
No error handling guidance. If fetch fails (timeout, 404, invalid URL), what should the LLM do? No recovery hints provided.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 30 | 2026-07-28+ | v2 |
Tool name 'fetch_web_content' is ambiguous. Does it fetch a full page, a snippet, search results, or metadata? Rename to 'search_web', 'get_webpage', or 'fetch_url' with clearer intent.