Loading ToolBench
Two tools with minimal documentation. Both have descriptions (10-20 chars baseline), but lack depth and context. Parameter descriptions are trivial ('Bob' is not a description). No output schemas documented. No error handling guidance. No parameter constraints or validation hints. Tool names are acceptable (verb_noun pattern) but descriptions fail to explain WHEN to use each tool or what they return. This server falls well below production baseline.
Get weather forecast for a location.
Provides a hello world response for a username
Parameter descriptions are placeholder text, not actual documentation. 'Bob' for the name parameter and generic coordinate descriptions do not explain constraints, formats, or valid ranges.
No output schema documented for either tool. LLMs cannot plan downstream calls or extract structured data when return types are unknown.
No error handling guidance. If NWS API fails or coordinates are invalid, the tool returns a generic string. LLMs cannot distinguish retryable errors from user-fixable ones.
get_forecast lacks numeric constraints. Latitude/longitude have no min/max bounds documented. LLMs may pass invalid coordinates (e.g., latitude > 90) without guidance.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
Tool descriptions are too short (10-20 chars). Baseline for A+ tools is 50-200 chars. Current descriptions lack context on WHEN to use each tool and what they return.