Loading ToolBench
This server has critical definition quality gaps. Both tools lack parameter descriptions entirely (all 'description': null in the schema). The 'sum' tool has a misleading description, it says 'Add two numbers' but the implementation multiplies them (a * b), creating a dangerous semantic mismatch. The 'housePrice' tool description is vague and lacks guidance on when/why to use it. No output schemas are documented. Parameter names lack type suffixes to disambiguate (e.g., 'area' should be 'area_sqft' or 'area_sqm'). No error handling or recovery guidance. No tool-level constraints (enums, ranges) are defined. The 'sum' tool itself violates the single-responsibility principle by existing alongside a domain-specific tool.
use number of bedrooms, no of bathrooms and location to get the price of houses
Add two numbers
CRITICAL: 'sum' tool description says 'Add two numbers' but implementation multiplies (a * b). This is a semantic mismatch that will cause agents to misuse the tool and produce wrong results.
All tool parameters lack descriptions (all 'description': null in schema). Parameter meanings are ambiguous, 'area' could be total area, buildable area, or land area; 'location' has no format guidance.
No output schemas documented. Agents cannot determine what fields 'housePrice' returns (float? currency object? with confidence intervals?). This forces agents to guess and wastes tokens on trial-and-error.
'sum' tool should not exist in a price-prediction domain. It violates single-responsibility and forces the agent to reason between a generic math tool and a domain-specific tool. Remove it or move it to a separate 'math' server.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 32 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 21 | - | v1 |
Parameter naming is not self-documenting. 'area' lacks units; should be 'area_sqft' or 'area_sqm'. 'location' accepts a string but has no guidance on format (city name? postal code? geographic coordinates?).
No error handling or recovery guidance. If 'location' is invalid (e.g., a location not in the model's training data), what does the tool return? The code catches the exception silently (bare except), so the agent gets an undefined result and cannot self-correct.
No parameter constraints or enums. 'bedrooms', 'bathrooms', and 'area' are integers with no min/max bounds. An agent could pass negative values, zero bedrooms, or absurdly large areas, all of which are invalid for house pricing but will not be rejected.
'housePrice' description is generic and lacks actionable guidance. It does not specify when to use this tool vs. a similar one, what the expected price range is, or what assumptions underlie the model (training data era, geographic scope, etc.).