Two tools with complete schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Tool names follow verb_noun convention (search_, discover_) which is good. Descriptions are adequate (150-200 chars, within baseline range). However, many optional parameters lack actionable constraints, e.g., 'page' has no min/max bounds, 'sort_by' has no enum of valid values, date parameters lack format validation guidance. No documented output schema structure. Error handling present but generic. No tool annotations (readOnlyHint, idempotentHint) despite both being READ_ONLY operations.

• Add explicit output schema documentation: 'Returns {results: [{id, name, biography, known_for_department, profile_path}], total_results, total_pages, page}' for search_person; similar for discover_films.
• Convert sort_by, with_release_type, with_watch_monetization_types to enum constraints in Zod schema. Replace description examples with formal enum values.
• Add min/max constraints to numeric parameters: page (1-1000), vote_average_gte/lte (0-10), vote_count_gte (0-unlimited), with_runtime_gte/lte (0-500).
• Add tool annotations via server.tool() options: {readOnlyHint: true, idempotentHint: true} to both tools so agents know they are safe to retry.
• Enhance error handling: categorize as 'retryable' (rate limit, timeout) vs 'user-fixable' (invalid query, missing token). Return 'Try search_person() with a simpler query' or 'Check ACCESS_TOKEN environment variable' instead of generic messages.