creating-dbt-models

What problem does it solve?

It solves the problem of reliably creating and updating dbt models without introducing schema mistakes or incorrect transformations by enforcing build-and-verify discipline.

Core Features & Use Cases

• Conventions-first model creation: Discovers naming, config, and SQL patterns from your existing project before writing new logic.
• End-to-end validation: Runs dbt build (not compile) and then verifies the resulting relations and sample data using dbt show.
• Correctness checks against requirements: Confirms grain, column names, NULL expectations, and calculation outputs against sample rows and schema.yml test expectations.
• Use case: When you need to add a new fact or intermediate model (or modify joins, transformations, or columns) based on task requirements or schema.yml specs, this ensures the model matches both structure and logic.

Quick Start

Use the creating-dbt-models skill to create or modify a dbt model following project conventions, run dbt build for verification, and then confirm the output with dbt show using the requested columns and grain.