Loading ToolBench
HTTP transport MCP server with FastMCP framework
Single tool 'add' has critical definition gaps. Tool itself lacks a description entirely (null/missing). No output schema documented. No error handling guidance. Naming is generic ('add') without context about what domain or operation. This server fails multiple hard scoring rules.
Tool 'add' has no description. LLMs cannot determine when or why to select this tool.
Both parameters (a, b) have null descriptions. Parameter descriptions are mandatory, LLMs cannot infer that 'a' and 'b' are operands for arithmetic addition.
No output schema documented. LLMs do not know the return type or structure (is it int? object? error details?).
Tool name 'add' is generic and lacks domain context. Does it add numbers, items to a list, users to a group, or something else?
No error handling or recovery guidance. If addition fails (e.g., integer overflow), the LLM has no actionable error message.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 33 | 2026-07-28+ | v2 |