build-data-spec

You generate dbt documentation by scanning your SQL and YAML definitions to identify analytics events, models, and columns, then producing a ready-to-use Markdown data spec for your repository.

A data spec is a Markdown reference documenting events, models, and columns across your data layer. You need one when starting a new dbt analysis, documenting a feature domain, or creating a reference for analysts.

Document dbt event models by reading the SQL and YAML definitions across your base, intermediate, and mart layers. The skill reads the codebase and writes a structured Markdown spec to your docs directory.

Yes, the skill works directly with existing dbt SQL and YAML files. It reads the event registry, macros, and model definitions from your repository to extract and document the relevant analytics columns.

Yes, you can create a reference data spec by running the skill against your entire dbt transforms repository. It outputs a Markdown file to your analytics docs directory for analysts and agents to use.

The best way to document a feature domain is to scan all related dbt models and event registries in your codebase, automatically extracting columns and definitions to generate a comprehensive Markdown spec.