Skill UI

This skill converts vague business questions or monitoring needs into a complete, structured, and implementation-ready dashboard specification. It guides users to define key metrics (KPIs), chart types, axes, filters, and optimal layout, providing data engineers and BI developers with all necessary details to build the dashboard without ambiguity.

Run arbitrary PySpark or pure Python code directly on Fabric Spark compute using the Livy API. These sessions are ephemeral, meaning no notebook artifacts are created or persisted, making it ideal for automated ETL processes, programmatic data transformations, or agent-driven compute workflows. Provides full read/write access to lakehouse Delta tables via Spark SQL.

This skill enables querying Delta Lake tables and raw files stored in OneLake using DuckDB. It supports two primary modes: local execution via CLI/Python for ad-hoc analysis, and in-notebook execution within Fabric, combining DuckDB's speed with notebook context. Ideal for data exploration, schema discovery, and data quality validation across the Fabric ecosystem.

This static analysis skill audits SQL queries across major data warehouses (BigQuery, Snowflake, Redshift, Postgres) for performance and cost anti-patterns (e.g., SELECT *, full scans, Cartesian joins). It provides a health score (0-100) and generates a prioritized, actionable cost-reduction plan, detailing the 'why' and the 'how' to fix expensive queries.