Six tools with complete schemas and descriptions. Naming follows verb_noun pattern (db2_lookup, db2_search, db2_validate_ids, db2_query, db2_tables, db2_describe). All tools are READ_ONLY with readOnlyHint annotations. Descriptions are adequate (100-200 chars) and explain WHAT and WHEN. Parameters have types and descriptions. However, output schemas are not documented, LLMs cannot predict response structure. Error handling is minimal (ValueError exceptions without recovery guidance). No pagination support despite tools returning potentially large result sets. Parameter constraints (e.g., limit range, SQL validation) lack explicit bounds in descriptions.

• Document output schemas for all tools. E.g., db2_lookup returns [{id: int, ...columns}], db2_search returns [{id: int, name: string, ...}]. Include field types and which fields are always present.
• Add pagination to db2_query and db2_search. Introduce offset/limit or cursor-based pagination. Return total_count and has_more flags. Update descriptions to state 'Returns up to N results; use offset to fetch additional pages.'
• Enhance error messages with recovery guidance. E.g., 'Unknown table "foo". Call db2_tables() to list available tables.' or 'No searchable column found. Call db2_describe(table) to see column names, then specify column explicitly.'
• Explicitly bound numeric parameters. Add to db2_search description: 'limit: integer, 1 - 1000 (default 25)'. Add to db2_query: 'Queries timeout after 30 seconds. Default LIMIT 1000 enforced if omitted.'
• Add a db2_query_explain tool or document query complexity constraints. Warn agents about expensive JOINs or aggregations on large tables. Alternatively, add a 'dry_run' parameter to estimate cost before execution.