This server has 14 tools with basic descriptions but critical gaps in schema documentation and parameter clarity. Tool names follow verb_noun convention (get_, execute_, add_) which is good. However, input schemas are inferred from C# type signatures rather than explicitly documented JSON Schema. Parameters like 'services' (IServiceProvider) and 'cancellationToken' are infrastructure concerns that should not appear in tool definitions exposed to LLMs. Most tools lack output schema documentation. Descriptions are present but generic (10-50 chars), below the 50-200 char baseline for LLM optimization. No enum constraints, no error recovery guidance, no pagination support despite tools like execute_reader that could return large result sets.

• Remove 'services' and 'cancellationToken' from tool parameter definitions. These are server-side concerns. Expose only user-facing inputs (schema, tableName, sql, parameters).
• Document output schemas for all tools. E.g., xbase_get_tables should return {tables: [{name: string, schema: string, rowCount: int}]}. This enables LLMs to plan downstream calls.
• Expand descriptions to 80-150 characters. Example: 'Get list of tables in the database' → 'Retrieve all tables in the specified schema. Use this first to explore available tables before querying. Returns table names and row counts.'
• Add pagination to xbase_execute_reader: accept 'limit' (default 100, max 1000) and 'offset' parameters. Return total_rows and has_more flags so LLMs can fetch additional pages.
• Add error recovery hints to execute tools. E.g., 'If SQL syntax error occurs, check column names with xbase_get_columns(). If permission denied, verify database credentials.'