Two tools with basic verb-noun naming (get_alerts, get_forecast). Both have short descriptions (34-40 chars, below the 50-200 char LLM-optimized baseline). Input schemas are present with types and descriptions, but output is unstructured plain text. No error recovery guidance, no pagination support, no structured response schema documented. Lacks tool annotations (readOnlyHint present in metadata but not in schema). No parameter validation constraints (e.g., state code format, lat/lon bounds). Error handling returns generic strings without actionable recovery steps.

• Expand descriptions to 80-150 chars. Example: 'get_alerts: Retrieve active weather alerts for a US state. Returns event type, severity, affected area, and instructions. Use when user asks about weather warnings or hazards.'
• Return structured JSON objects instead of formatted strings. get_alerts should return {alerts: [{event, area, severity, description, instructions}], count, state}. get_forecast should return {periods: [{name, temperature, windSpeed, windDirection, forecast}], location, generatedTime}.
• Add input validation: state as enum of valid US state codes; latitude/longitude with min/max bounds and description 'Must be valid coordinates: latitude -90 to 90, longitude -180 to 180'.
• Enhance error messages with recovery hints. E.g., 'Invalid state code: got "XX". Valid codes are 2-letter US state abbreviations (CA, NY, TX, etc.)' or 'Location outside NWS coverage area. Try a different latitude/longitude within the continental US.'
• Add limit and offset parameters to get_forecast (e.g., limit: 1-14, default 5) and document in description: 'Returns up to 14 forecast periods. Use limit parameter to control result size.'