This is a minimal example server for testing the mcp-test framework itself, not a production MCP server. Both tools have basic schemas and descriptions, but critical gaps prevent real-world use: parameter descriptions are missing or null (add tool has no param descriptions), descriptions are too short to guide LLM selection, no error recovery guidance, no output schema documentation, and no tool annotations. The server is intentionally simplistic for testing purposes.
Add two integers.
Return the input message unchanged.
add tool: parameter descriptions are null/missing. The 'a' and 'b' parameters have no descriptions, violating the requirement that every parameter must explain what it controls. LLMs cannot infer that these are integers to be summed.
Tool descriptions are too short (10-30 chars). 'Return the input message unchanged' and 'Add two integers' lack context for LLM selection. Descriptions should be 50-200 chars and answer: what does it do, when to use it, what does it return?
No output schema documented. Tools return text content but LLMs have no formal schema describing the response structure. This forces LLMs to guess at field names and types.
No error recovery guidance. The add tool silently defaults missing parameters to 0 (args.get('a', 0)). If an LLM passes invalid input, there is no actionable error message to guide correction.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 37 | 2026-07-28+ | v2 |
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Both tools are read-only but this is not declared in the schema, forcing LLMs to infer safety from names alone.