Two tools with basic verb-noun naming (get_alerts, get_forecast). Both have short descriptions (30-40 chars, below the 50-200 char baseline for LLM-optimized descriptions). Input schemas are present with typed parameters but lack descriptions for all fields. Output is unstructured plain text rather than structured JSON. No error handling guidance, no pagination support, no output schema documentation. Naming is clear but descriptions are too terse to guide LLM tool selection effectively.

• Expand tool descriptions to 80-150 chars. Example: 'Get active weather alerts for a US state. Returns event type, severity, area, and instructions. Use this to check for warnings before get_forecast.'
• Add parameter descriptions: state → 'US state code (2 letters, e.g. CA, NY, TX)', latitude → 'Latitude coordinate (-90 to 90)', longitude → 'Longitude coordinate (-180 to 180)'
• Return structured JSON instead of plain text. Example: {alerts: [{event, area, severity, description, instructions}]} or {periods: [{name, temperature, windSpeed, windDirection, forecast}]}
• Add input validation: validate state against 50 US state codes, clamp lat/lon to valid ranges, return actionable errors like 'Invalid state code: got "XX", must be 2-letter US state code (e.g. CA, NY, TX)'
• Document output schema in tool descriptions and add error recovery hints: 'If location not found, try nearby coordinates. If state code invalid, use 2-letter abbreviation (CA not California).'

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry