Build and validate dbt data models
Start with the target table, work back to its sources and validate transformations against data. dbt Labs connects SQL, model dependencies and tests.
711 · GitHubdbt-labs/dbt-agent-skillsEditorial selection · source reviewedAuthor: dbt Labs Apache-2.0 Revision 2116bc1
When to use it
Build a sales mart with an explicit row grain
Investigate duplicates introduced by joins
Assess the impact of changing an existing model
What you’ll get
dbt models, tests and documented checks of input and output data.
For example
Design a daily sales mart. Define the row grain, inspect existing models and sources, then add SQL and tests for refunds and duplicates.
Original, requirements and limits
Original skill: dbt Labs. Editorial description: kossolapov.com.
- Agent
- An agent that can read Agent Skills instructions; tool access depends on the host
- Tools
- dbt CLI and an existing dbt project with the appropriate warehouse adapter
- dbt show access to test data; optional dbt MCP integration
- For breaking changes: companion working-with-dbt-mesh, included in the source file list
- Access
- Authorized test warehouse profile or credentials
- Possible costs
- Warehouse queries and optional dbt platform services may be billed
- Limits
- Requires real schema and data access; does not infer business definitions reliably from names alone.
- Breaking column changes require the companion working-with-dbt-mesh skill and a migration plan.
- Use selective runs and bounded exploration to control warehouse cost.
- Editorial selection based on source inspection, not a comparative performance certification.