Two tools with basic verb-noun naming (get_flights, get_flight_details). Both have descriptions (READ_ONLY risk classification visible). Input schemas present with typed parameters and descriptions. However, output schemas are NOT documented in the source code, critical gap for LLM planning. Parameter descriptions are minimal (e.g., 'The flight ID' lacks format/constraint guidance). No error handling guidance visible. Missing pagination support despite flights being a list result. No tool annotations (readOnlyHint, idempotentHint) despite READ_ONLY risk classification.

• Document output schemas for both tools. Example for get_flights: 'Returns array of {id: string, airline_name: string, departure_time: ISO8601, arrival_time: ISO8601, price: number, available_seats: number}'. For get_flight_details: 'Returns {airline_name, departure_airport, arrival_airport, departure_time, arrival_time, duration, price, available_seats}'.
• Add pagination to get_flights: include limit (default 20, max 100) and offset parameters; return total_count and has_more in response.
• Enhance parameter descriptions: 'departure: IATA airport code (3 letters, e.g. JFK) or city name (e.g. New York)' and 'flight_id: Unique flight identifier returned by get_flights (alphanumeric, 6-8 characters)'.
• Add tool annotations in fastmcp registration: mark both tools with readOnlyHint=true to signal they are safe, non-destructive queries.
• Add error handling: document what happens if flight_id not found ('Flight not found. Try get_flights() to search for available flights.') and if no flights match search criteria ('No flights found for this route/date. Try alternative dates or nearby airports.').