SQLMesh

Build transformation pipelines with SQLMesh. Writes MODEL blocks and Python models, chooses incremental kinds, runs plan/apply/audit, manages virtual environments and snapshots, and migrates dbt projects.

This skill makes Claude fluent in SQLMesh. It writes SQL MODEL blocks and Python @model decorators, picks the right model kind (incremental by time or unique key, SCD type 2, view, seed), runs plan/apply/audit/test, manages virtual environments, snapshots, and state, configures gateways and hooks, and migrates dbt projects. SQLMesh moves fast, so the skill fetches canonical docs before non-trivial work.

When to use

Use when writing SQLMesh models, choosing an incremental model kind, running plan/apply/audit, managing snapshots and virtual environments, or migrating a dbt project to SQLMesh.

Examples

Incremental model

Write a time-range model

Write a SQLMesh INCREMENTAL_BY_TIME_RANGE model that builds daily revenue from raw orders, partitioned by order date

Plan and apply

Promote changes safely

Walk me through sqlmesh plan and apply for a change to my staging model, explaining the snapshots it creates

Migrate from dbt

Convert a project

Convert my dbt incremental model into an equivalent SQLMesh model and explain the differences