Server has 4 tools with basic schemas and descriptions, but significant gaps prevent higher scores. All tools have descriptions (10-80 chars, mostly adequate) and input schemas with type definitions. However, parameter descriptions are sparse or missing entirely. list_locations and list_media lack any parameter descriptions (empty properties). create_campaign has good parameter coverage but missing output schema documentation. No error recovery guidance, no pagination support for list operations, and no tool annotations (readOnlyHint/destructiveHint). Naming is clear and verb-based. Overall structure is functional but falls short of production-grade quality.

• Add pagination to list_locations and list_media: include limit (default 20, max 100), offset/cursor, and total_count in responses. Document in tool descriptions: 'Returns up to 20 locations per call. Use offset to fetch additional results.'
• Document output schemas for all tools. Example for list_locations: 'Returns {locations: [{id, name, status, screen_count}], total_count, offset}'. This enables LLMs to chain calls (e.g., extract location_id for create_campaign).
• Enhance error messages with recovery guidance. Replace 'Error executing tool: ...' with 'Invalid screen_id: "xyz" not found. Call list_locations() to see available screen IDs.' Categorize errors as retryable (network timeout) vs. user-fixable (invalid input).
• Add tool annotations to the server.setRequestHandler response. Mark create_campaign with destructiveHint: true and idempotentHint: false. Mark list_* with readOnlyHint: true. This signals to agents which operations are safe to retry.
• Add validation and constraints to create_campaign date parameters. Validate ISO 8601 format, ensure start_date < end_date, and return clear error: 'Invalid start_date: must be ISO 8601 (e.g., 2024-01-15T10:00:00Z) and before end_date.' Add minLength/maxLength to name parameter.