Duckdb Sql

The duckdb-sql skill enables AI agents to execute SQL queries directly against DuckDB, an in-process analytical database. It allows agents to load data from various file formats and perform complex calculations without requiring a persistent server setup.

When to use it

• Analyzing local CSV, JSON, or Parquet files for immediate insights.
• Performing ad-hoc exploratory data analysis on small-to-medium datasets.
• Running SQL queries that require zero-latency execution within a single process.
• Generating reports or aggregating data without managing external database infrastructure.

Key capabilities

• Executes standard SQL queries against DuckDB instances.
• Loads data from multiple file formats including CSV, JSON, and Parquet.
• Operates entirely in-memory or with local file access, eliminating server dependencies.
• Supports complex aggregation, filtering, and joining operations natively.

Example prompts

• "Load the sales_data.csv file and calculate the total revenue by region."
• "Query the transactions table to find the top 10 customers by spending in Q4."
• "Join the users and orders tables to count active users who have placed more than five orders."

Tips & gotchas

Ensure your AI agent has read access to the local files or database paths you intend to query, as DuckDB operates on local resources. Be mindful of memory usage when loading very large datasets directly into the in-process engine.