Loading ToolBench
MCP server with OAuth 2.0 authentication provider running on Cloudflare Workers, demonstrating remote authorization flows
Single tool 'add' has a basic schema with typed parameters (number, number) and minimal descriptions. Tool name lacks action verb (should be 'add_numbers'). Description is extremely terse (5 words, well below 20-char threshold). No output schema documented. No error handling guidance. Parameter descriptions are present but generic. This is a toy/demo implementation with significant gaps vs. production baselines.
Adds two numbers together
Tool name 'add' lacks action verb. Should be 'add_numbers' or similar to match verb_noun convention. LLMs rely on names to infer intent before reading descriptions.
Tool description 'Adds two numbers together' is only 28 characters, below the 50-char minimum for LLM clarity. Should explain WHEN to use it, what it returns, and any constraints (e.g., numeric range limits).
No output schema documented. LLMs cannot plan downstream calls or extract results without knowing the response structure. Should document that output is a number (or text representation of sum).
Parameter descriptions ('First number to add', 'Second number to add') lack constraints. No mention of numeric range, precision, or handling of edge cases (infinity, NaN, very large numbers).
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | 2026-07-28+ | v2 |
No error handling guidance. What happens if inputs are non-numeric, null, or undefined? LLM receives no recovery path if the call fails.