Four tools with comprehensive parameter schemas and clear descriptions. All tools have proper verb-noun naming (search_*, get_*). Descriptions are detailed (150-250 chars) and explain WHAT and WHEN. Input schemas are well-structured with types and descriptions for most parameters. However, output schemas are not documented, LLMs cannot predict response structure for downstream chaining. Error handling guidance is absent. Parameters lack enum constraints where applicable (e.g., cabin_class, max_stops, sort_by accept free-form strings instead of enums). No tool annotations (readOnlyHint, idempotentHint) despite all being read-only operations.

• Document output schemas for all tools. Specify fields (flight_id, price, departure_time, etc.), types, and pagination structure (total_count, next_cursor). This enables LLMs to extract data and chain tools.
• Convert free-form string parameters to enums. Replace cabin_class: string with enum [ECONOMY, PREMIUM_ECONOMY, BUSINESS, FIRST]. Replace max_stops with enum [ANY, NON_STOP, ONE_STOP, TWO_PLUS_STOPS]. Replace sort_by with enum [CHEAPEST, DURATION, DEPARTURE_TIME, ARRIVAL_TIME].
• Add numeric bounds to all integer parameters. E.g., passengers: {type: integer, minimum: 1, maximum: 9}, children: {minimum: 0, maximum: 8}, checked_bags: {minimum: 0, maximum: 2}.
• Add tool annotations. Mark all four tools with readOnlyHint: true since they perform no mutations. This signals to agents that retries are safe.
• Enhance error messages with recovery guidance. E.g., 'Invalid airport code "XYZ". Try search_airports(query="your city") to find valid codes.' This guides agents to self-correct.