Postgres SQL Tool

​

What it does

​

Key features

• Execute any SQL query (SELECT, INSERT, UPDATE, DELETE)
• Results are truncated to a 30,000-token budget, so a wide query returns as many whole rows as fit
• Clean, structured results that agents can easily parse
• Built-in error handling for common SQL issues

​

Parameters

​

Common use cases

​

Data analysis

SELECT 
  DATE_TRUNC('month', order_date) as month,
  COUNT(*) as total_orders,
  SUM(total_amount) as revenue
FROM orders 
WHERE order_date >= '2024-01-01'
GROUP BY month
ORDER BY month DESC;

​

User lookups

SELECT user_id, email, last_login, account_status
FROM users 
WHERE email = 'user@example.com';

​

Performance monitoring

SELECT COUNT(*) as active_sessions
FROM pg_stat_activity 
WHERE state = 'active';

​

Security best practices

• Use read-only database users when possible
• Avoid querying sensitive data (passwords, API keys, PII)
• Use WHERE clauses to limit data exposure
• Monitor agent queries regularly

​

Troubleshooting

• Check that your Postgres integration is configured and connected
• Verify database server is running and accessible

• Ensure your database user has SELECT permissions
• Check that the table exists and is accessible

• Validate your SQL syntax
• Check table and column names for typos

​

Related tools

• Semantic Search Postgres (semantic_search_pg) - Search using natural language
• Keyword Search Postgres (keyword_search_pg) - Full-text search across content

​

Tool Metadata

• Name: postgres_sql
• Display Name: Postgres SQL
• Description: Execute a SQL query on Postgres
• Required Integration: Postgres
• Result limit: 30,000 tokens; rows past the budget are dropped and the response is flagged as truncated