What problem does it solve? It guides AI agents through dbt analytics engineering work—building models, writing tests, debugging errors, and assessing downstream impact—while enforcing best practices like DRY principles, data validation with dbt show, and avoiding breaking changes to models with consumers. ## Core Features & Use Cases - Model Building & Planning: Work backwards from desired output, mock upstream models, and validate results iteratively with dbt show before finalizing SQL. - Data Discovery & Testing: Profile unfamiliar sources, document findings, and write prioritized data tests (unique, not_null, relationships, accepted_values) based on actual discovery results. - Impact Assessment & Debugging: Evaluate downstream dependencies before changes using dbt ls selectors or MCP lineage tools, and classify/resolve parsing, compilation, and data errors. - Use Case: When asked to add a revenue metric to a dbt project, the skill reads existing model YAML docs, profiles source data with dbt show, plans the transformation, writes the model with ref() calls, adds targeted tests, and builds only affected downstream models with state:modified+. ## Quick Start Ask the agent to build a new dbt staging model for a specific source table and validate the output with dbt show.
