Single tool 'query' has a clear verb-noun name and basic schema, but lacks depth in descriptions and parameter guidance. Tool description is 95 chars (below baseline 194), parameter description is minimal (23 chars), and no output schema is documented. No error recovery guidance. Security validation (read-only check) is present but not reflected in tool annotations. Missing idempotent/readOnly hints.

• Expand tool description to 150 - 250 chars: 'Execute read-only SQL queries (SELECT, WITH, SHOW, EXPLAIN) against the AWS PostgreSQL database. Returns structured result sets with row count, field metadata, and data. Use this to retrieve data; for schema discovery, use the resources API to browse tables and schemas first.'
• Enhance 'sql' parameter description: 'SQL query string (SELECT, WITH, SHOW, DESCRIBE, EXPLAIN only). Must not contain INSERT, UPDATE, DELETE, DROP, CREATE, ALTER, TRUNCATE, GRANT, or REVOKE. Example: SELECT * FROM public.users WHERE id = $1'
• Document output schema in tool definition: { type: 'object', properties: { rowCount: { type: 'integer' }, rows: { type: 'array', items: { type: 'object' } }, fields: { type: 'array', items: { type: 'object', properties: { name: { type: 'string' }, dataTypeID: { type: 'integer' } } } } } }
• Add readOnlyHint: true to tool definition to signal safety and enable client-side optimizations.
• Improve error messages: 'Query rejected: only SELECT, WITH, SHOW, DESCRIBE, EXPLAIN allowed. Your query starts with [detected keyword]. Try rephrasing as a SELECT statement.' Include the detected keyword to help LLM self-correct.