Loading ToolBench
A simple MCP server that provides basic mathematical operations
Single tool 'add' has a clear, concise description (24 chars) and properly typed integer parameters with descriptions. However, the tool lacks critical production-grade patterns: no error handling guidance, no output schema documentation, no idempotency hints, and no recovery paths. The description is minimal but adequate. Parameters are well-typed but lack constraints (e.g., min/max bounds). No composition patterns, no pagination, no security considerations. This is a toy example, not production-ready.
Add two numbers together
No output schema documented. LLMs cannot infer return type or plan downstream operations.
No error handling or recovery guidance. Tool provides no actionable error messages or retry logic.
Numeric parameters lack bounds. No min/max constraints on 'a' and 'b', LLMs may pass absurd values.
No idempotency or side-effect declaration. Tool description does not clarify if it is safe to retry.
Tool name 'add' is generic and lacks verb clarity. Consider 'calculate_sum' or 'add_integers' for better LLM intent inference.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 52 | 2026-07-28+ | v2 |