MCP server that serves a grocery store widget and exposes a tool to search for grocery items. Supports both OpenAI and MCP Apps transports with SSE protocol.
Single tool 'search-groceries' has a clear verb-noun name and basic schema with typed input. However, the description is minimal (89 chars, below the 194-char baseline), lacks context on when/why to use it, and provides no guidance on output structure, pagination, or error handling. No documented return schema visible. Tool appears functional but lacks production-grade documentation and composition guidance.
Takes a user query and returns a list of matching grocery items to display in the carousel.
Tool description is too brief (89 chars vs 194-char baseline). Lacks WHEN to use, WHAT it returns, and error recovery guidance. LLMs cannot determine selection criteria or downstream expectations.
No documented output schema. LLMs cannot plan downstream tool calls or extract structured data without knowing return fields, types, and pagination behavior.
No error handling guidance. Tool lacks recovery hints for invalid queries, empty results, or service failures. LLMs receive no actionable error messages.
Input schema lacks constraints. 'query' parameter has no length limits, format hints, or examples. LLMs may pass empty strings or excessively long queries.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2025-06-18+ | v2 |
No pagination or result-limiting guidance. Tool description does not specify max results, whether pagination is supported, or how to handle large result sets.