dbt-migration-validation_skill

Overview

This skill provides comprehensive validation for dbt models and schema YAML files during migrations. It enforces validation rules, detects common anti-patterns, and offers auto-fix suggestions to keep models Snowflake-compatible and production-ready. It integrates with validation hooks to run checks automatically on edits and writes.

How this skill works

The skill inspects schema YAML and SQL model files to apply a ruleset for documentation, naming, tests, CTE structure, and Snowflake compatibility. It detects missing descriptions, absent primary/foreign key tests, SELECT * usage, hardcoded table references, and platform-specific syntax, then recommends concrete fixes. Validation runs via configured hooks and returns non-zero exit codes for errors while reporting warnings.

When to use it

• During model migration from other platforms to Snowflake to catch incompatible syntax and breaking changes.
• When reviewing pull requests to ensure models follow naming, testing, and documentation standards.
• While diagnosing validation hook failures triggered on file edits or writes.
• When configuring or enforcing team validation thresholds and automated checks.
• To generate targeted auto-fix suggestions for common dbt anti-patterns.

Best practices

• Always include model and column descriptions in schema YAML to improve docs and ownership.
• Declare dbt_constraints.primary_key and relationship tests for key-like column names.
• Use ref() and source() for all table references to preserve lineage and environment portability.
• List explicit columns in final SELECTs; avoid SELECT * except from a final CTE that lists columns.
• Follow layer-specific naming prefixes (stg_, int_, dim_, fct_) to make model intent clear.
• Prefer standard CTE patterns and a config block with materialization for readability and consistency.

Example use cases

• Validate a migrated SQL model to replace TOP/ISNULL/GETDATE() with Snowflake equivalents and add a migration header.
• Scan schema YAML to add missing model descriptions and data_type declarations before releasing documentation.
• Fail CI on critical errors like missing primary key tests or hardcoded table references.
• Automatically run validators on file save to get immediate auto-fix suggestions for CTE structure or naming issues.
• Generate a remediation checklist for large migration batches highlighting platform-specific fixes and breaking changes.