Loading ToolBench
GitHub: Tetsuya-Minase/mcp-server-sample
Two tools with minimal descriptions (both under 30 chars), no parameter descriptions, and no documented output schemas. Tool 1 has empty input schema ({}); Tool 2 has a single enum parameter with description. Neither tool documents what it returns or when to use it. Naming is verb-based but descriptions are trivial. No error handling guidance. This server falls well below production baseline (194 char avg description, 100% param coverage in A+ tools).
get tminasen infomation
get tminasen's skill infomation
Tool 1 (get-tminasen-info) has empty input schema {} with no description. Description is 24 chars ('get tminasen infomation'), below 20-char threshold, capped at 15.
No parameter descriptions. Tool 2's 'genre' parameter has a description in code ('skill genre'), but Tool 1 has no parameters to describe. Tool 2's description is only 32 chars ('get tminasen's skill infomation'), below baseline 194 chars. LLMs cannot determine when to call these tools or what they return.
No output schema documented. Both tools return hardcoded text responses, but the schema (content array with type and text fields) is not declared in tool definitions. LLMs cannot plan downstream calls or extract structured data.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 41 | 2026-07-28+ | v2 |
No error handling or recovery guidance. If a tool fails or returns unexpected data, the LLM has no actionable next step. No validation of inputs (e.g., genre enum is enforced by Zod but not documented in description).
Typo in description: 'infomation' should be 'information'. Unprofessional and suggests lack of review.