Single tool with adequate naming but significant gaps in schema clarity, output documentation, and error handling. The tool is explicitly registered via @mcp.tool() decorator with visible implementation, so not capped at 50. However, the input schema shows only basic structure with minimal type information, the return type is undocumented (free-form string), error messages are generic, and no guidance is provided for recovery. The description is adequate but lacks information about prerequisites (requiring external APIs to be accessible) and does not explain when or why an LLM should call this tool vs alternatives. Baseline calibration: most community servers score 45-55; this server is below average due to missing output schema documentation, lack of error recovery guidance, and no parameter validation constraints.
Get weather information for a city.
Output schema not documented. Tool returns a free-form string with embedded newlines, but LLM does not know the structure, fields, or how to parse it. Agents cannot extract specific fields (temperature, wind speed) for downstream reasoning.
Input parameter lacks format constraints and validation guidance. 'city' is documented as accepting string like 'San Francisco, CA' but no constraint on length, character set, or format is stated. No enum or pattern validation, LLM could pass invalid city names, and error messages are generic ('Could not find coordinates').
Error messages lack recovery guidance. When geocoding fails, tool returns 'Could not find coordinates for city: {city}' with no suggestion for next steps. LLM has no guidance: should it retry? Try a different city format? Consult a discovery tool?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 45 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 37 | - | v1 |
Tool description does not explain prerequisites or external dependencies. The tool calls two external APIs (OpenStreetMap Nominatim, National Weather Service) with 30s and 10s timeouts respectively. If either is unavailable, tool fails silently with generic error. Description should mention this and hint at retry strategy.
No pagination or result limits documented. Tool returns a single forecast period, but the API response can contain multiple periods. Tool silently ignores all but the first, no documentation of this behavior, and no parameter to request extended forecast.