What databases are supported?

It supports databases configured in your warehouse.yml file, typically including major cloud data warehouses like Snowflake, BigQuery, and Redshift.

What is the primary purpose of the /data:init command?

It initializes your data environment by scanning your warehouse and local code to create a schema reference file, allowing Claude to understand your data model without constant database queries.

Does this skill work with dbt?

Yes, it specifically searches for dbt model YAML files and schema definitions to extract column descriptions and primary key information.

How often should I run the initialization skill?

You should run it once per project and whenever your warehouse schema changes, such as after deploying new dbt models or migrating tables.

Data Warehouse Schema Discovery

Name: Data Warehouse Schema Discovery
Author: astronomer

byastronomer

•

Database Management

Automates data warehouse schema discovery and generates comprehensive documentation enriched with codebase context for AI-assisted data engineering.

About

The Data Warehouse Schema Discovery skill (init) streamlines data engineering workflows by scanning your configured databases and local codebase to create a centralized schema reference. It intelligently parses dbt models and SQL frontmatter to enrich technical metadata with business context, identifies high-volume tables to prevent inefficient queries, and generates a version-controllable documentation file. By caching these relationships, it enables Claude to perform instant concept-to-table lookups, significantly increasing the accuracy of generated SQL and data analysis tasks.

Key Features

Identification of large tables (>100M rows) for safer querying
77 GitHub stars
Context enrichment from dbt models and declarative SQL frontmatter
Categorical value family discovery for precise filtering
Automated discovery of databases, schemas, and table metadata
Generation of a shareable .astro/warehouse.md reference file

Use Cases

Initializing a new data project for AI-assisted SQL development
Refreshing schema documentation after running data pipeline migrations
Improving LLM accuracy for complex queries by providing local schema context

About

Key Features

Identification of large tables (>100M rows) for safer querying
77 GitHub stars
Context enrichment from dbt models and declarative SQL frontmatter
Categorical value family discovery for precise filtering
Automated discovery of databases, schemas, and table metadata
Generation of a shareable .astro/warehouse.md reference file

Use Cases

Initializing a new data project for AI-assisted SQL development
Refreshing schema documentation after running data pipeline migrations
Improving LLM accuracy for complex queries by providing local schema context