Integrated MCP server providing math calculation and weather information tools
Both tools have critical definition gaps. 'calculate' has a trivial 34-char description that violates the 10-1024 char guideline and provides no context on when to use it or what it returns. 'get_weather' has a 27-char description with a grammatical error ('Get the weather location' is unclear). Neither tool documents output schemas, return types, or error handling. Parameter descriptions exist but are minimal. No tool annotations, no error guidance, no security considerations. The math tool uses eval() without input validation, a critical security flaw. Neither tool follows the verb_noun naming convention strongly (calculate is acceptable, but get_weather is generic). No pagination, no structured output documentation, no recovery guidance.
Evaluate math expression like '(3+5)*12'.
Get the weather location.
Missing output schema documentation. Neither tool documents what fields are returned or their types. LLMs cannot plan downstream calls or extract data reliably.
Descriptions are too short and lack context. 'calculate' (34 chars) and 'get_weather' (27 chars) do not explain WHEN to use the tool, WHAT it returns, or any prerequisites. Baseline is 194 chars average.
Security vulnerability: calculate() uses eval() without input validation. Agents can inject arbitrary Python code. Requires immediate sanitization or expression parser replacement.
No error handling or recovery guidance. Tools do not document what errors can occur, whether they are retryable, or what the LLM should do next.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 37 | 2026-07-28+ | v2 |
get_weather returns a hardcoded string ('It's always raining in California') regardless of input. Tool is non-functional and does not actually fetch weather data.