Two tools with basic verb-noun naming (get_alerts, get_forecast). Both have short descriptions (40-50 chars) and documented input schemas with type and description fields. However, descriptions lack context on WHEN to use each tool, prerequisites, or what the output structure contains. No output schema documentation. Error handling returns plain strings without recovery guidance. Parameters are minimally described (state: 'Two-letter US state code' is adequate but latitude/longitude lack range constraints). No pagination, no structured output, no error categorization.

• Expand tool descriptions to 150-250 chars. Include: what the tool does, when to call it (e.g., 'Use this to check active severe weather warnings'), what it returns (e.g., 'Returns a list of active alerts with event type, severity, and instructions'), and any prerequisites (e.g., 'Requires a valid two-letter US state code').
• Document output schema explicitly. For get_alerts, specify: returns string with formatted alerts or error message. For get_forecast, specify: returns string with 5 forecast periods (name, temperature, wind, detailed forecast). Consider returning structured JSON instead of formatted strings to enable LLM parsing.
• Add parameter constraints. For state: add enum of valid US state codes or pattern validation. For latitude/longitude: document valid ranges (-90 to 90 for lat, -180 to 180 for lon) in parameter descriptions.
• Implement error categorization and recovery hints. Instead of 'Unable to fetch alerts', return: 'Error: Invalid state code. Valid codes are: CA, NY, TX, ... (all 50 US states). Try again with a valid code.' or 'Error: API timeout. Retry in 5 seconds.'
• Add pagination support. For get_alerts, accept optional limit (default 10, max 50) and offset parameters. Return total count and next_offset. For get_forecast, document why only 5 periods are returned and offer a limit parameter.