Loading ToolBench
Single tool 'meshi-doko' has critical schema and naming defects. Schema is malformed (inputSchema.properties.inputs is an object with LOCATION/BUDGET as direct properties, not a proper nested object structure). Tool name lacks action verb, 'meshi-doko' is opaque Japanese slang, not verb_noun pattern. Description is minimal (8 words, 54 chars). Input schema structure is broken: 'user' is a string literal, not a property definition. Parameters LOCATION and BUDGET lack type declarations in the schema. No output schema documented. No error recovery guidance.
Send a message to Dify chat completion API
Input schema is malformed. 'inputs' property structure is invalid JSON Schema, LOCATION and BUDGET are direct object properties without type declarations. 'user' is a string literal, not a property object. This will fail schema validation.
Tool name 'meshi-doko' does not follow verb_noun convention and is opaque to LLMs. Should be 'send_restaurant_query' or 'search_restaurants' to clearly signal the action.
No output schema documented. LLMs cannot plan downstream calls or extract structured data from the response. Tool returns JSON.stringify(response.data), unstructured text.
Parameters LOCATION and BUDGET lack descriptions and type constraints. No enum, format, or validation guidance. LLMs cannot infer valid values.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 24 | 2026-07-28+ | v2 |
Error handling is minimal. Catch block logs error and re-throws without recovery guidance. LLM receives raw exception, cannot determine if retryable, user-fixable, or fatal.