Loading ToolBench
A Next.js-based MCP server using Vercel's MCP adapter
Single tool with minimal definition quality. Tool name 'dice_roll' lacks action verb prefix (should be 'roll_dice'). Description is in Japanese and only 28 chars, below the 34-char p10 baseline. Input schema is present with proper constraints (min/max), but parameter description is also in Japanese. No output schema documented. No error handling guidance. Tool is trivial (random number generation) but definition patterns are weak.
サイコロを降った結果を返します
Tool name lacks action verb. 'dice_roll' should be 'roll_dice' to follow verb_noun convention and help LLMs parse intent from the name alone.
Description is only 28 characters and in Japanese. Baseline is 34 - 392 chars. English description should explain WHAT the tool does, WHEN to use it, and what it returns.
No output schema documented. LLMs cannot plan downstream calls or extract data without knowing the response structure (e.g., is result an integer, object with 'value' field, or text?).
Parameter description 'サイコロの面の数' is in Japanese. All descriptions must be in English for LLM comprehension and international usability.
No error handling guidance. If sides parameter is invalid or out of range, no recovery hint is provided to the LLM.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 41 | 2026-07-28+ | v2 |