A test MCP server implementation in Java that provides tools for listing and searching items by category
Two tools with minimal descriptions (both under 50 chars) and incomplete parameter documentation. Input schemas are present but lack parameter descriptions, critical for LLM reasoning. No output schemas documented. No error handling guidance. Tool names follow verb_noun convention (get_, search_) but descriptions are too terse to guide LLM selection. Parameters lack type constraints, ranges, and format guidance. No indication of idempotency, side effects, or when to use each tool vs. the other.
Get a list of items based on category
Search items by title
Parameter descriptions missing. 'category' and 'query' parameters have no descriptions explaining what values are valid, expected format, or constraints. LLMs cannot infer whether 'category' accepts free text or only 'fruits'/'vegetables'/'all'.
Tool descriptions too short (both ~35 chars). Lack context for LLM selection. 'Get a list of items based on category' does not explain when to use get_list vs search_items, what data structure is returned, or whether results are paginated.
No output schema documented. Tool responses return unstructured text (JSON-like strings embedded in TextContent). LLMs cannot reliably parse or chain results. No indication of pagination, total count, or result structure.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 37 | <=2025-11-25 | v2 |
No error handling guidance. 'No category provided' error for get_list does not suggest recovery steps (e.g., 'Call get_list with category=all to see available options'). LLM has no path forward on failure.
Parameter constraints not declared. 'category' parameter should be an enum (fruits|vegetables|all) to prevent hallucinated values. 'query' parameter lacks length/format constraints.