Four tools with comprehensive schemas and detailed descriptions. All tools have clear verb-noun naming (search_, get_) and well-documented parameters with types. Descriptions are 50-200 chars, meeting LLM-optimization baselines. However, output schemas are not explicitly documented in the source code provided, and error handling guidance is minimal. Tools follow composition patterns well (search → get metadata → get data chain), but lack explicit error recovery hints and confirmation patterns for data retrieval operations.

• Document output schemas for all four tools. Specify field names, types, and structure. Example: 'Returns array of {indicator_id: string, database_id: string, title: string, description: string, last_updated: ISO8601}'. Include pagination metadata (total_count, has_more).
• Add error recovery guidance to each tool description. Example: 'If indicator_id not found, call data360_search_indicators() to find valid IDs. If API timeout occurs, retry after 5 seconds.'
• Enforce mutual exclusivity in data360_search_indicators: document that exactly ONE of query, queries, or query_groups must be provided. Consider adding a oneOf constraint in the schema or a validation error message.
• Document pagination behavior explicitly. State: 'Returns up to limit results. If result count equals limit, more results may exist. Use offset to fetch next page. Call with offset=0 to restart.'
• Clarify year parameter semantics in data360_get_data. Document: 'year is a shorthand for start_year=year, end_year=year. Do not pass year alongside start_year or end_year.' Add validation to reject conflicting parameters with a clear error message.