Loading ToolBench
A remote MCP server providing calculator tools without authentication, deployed on Cloudflare Workers
Two tools with minimal descriptions and basic schemas. 'add' (9 chars) and 'calculate' (10 chars) descriptions are far below the 50-200 char baseline for LLM-optimized tools. Both tools lack parameter descriptions entirely, critical gap per pattern:tool-description. Schemas are present but sparse. No error handling guidance, no output schema documentation, no composition hints. Tools are simple arithmetic operations with no real-world complexity, but even simple tools need proper documentation for agent selection and error recovery.
Simple addition tool
Calculator tool with multiple operations
Tool descriptions critically short (9-10 chars). Current descriptions lack context for agent selection.
No parameter descriptions. 'a', 'b', and 'operation' parameters lack explanations. LLMs cannot infer meaning from names alone, 'a' and 'b' are ambiguous without context.
No output schema documented. Callers cannot predict response structure. Tools return text content but no schema tells agents what fields to expect.
Divide-by-zero error returns plain text 'Error: Cannot divide by zero' instead of structured error with recovery guidance. Per pattern:recovery-guide, errors should tell LLM what to do next.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | 2026-07-28+ | v2 |
No tool annotations (readOnlyHint, idempotentHint). Both tools are read-only and idempotent but this is not declared in schema, forcing LLMs to infer safety properties.