Initial commit

commands/dataform-deploy.md

@@ -0,0 +1,21 @@
+---
+description: Deploy tested Dataform table to production
+---
+
+You are deploying a Dataform table to production using best practices from the dataform-engineering-fundamentals skill.
+
+**Workflow:**
+
+1. Invoke the dataform-engineering-fundamentals skill
+2. Ask the user which table they want to deploy
+3. **Pre-deployment verification:**
+   - Confirm the table has been tested in dev environment
+   - Verify all tests are passing
+   - Check that documentation (columns: {}) is complete
+4. **Deployment:**
+   - Run `dataform run --dry-run --actions <table_name>` (production dry-run)
+   - If successful, run `dataform run --actions <table_name>` (production execution)
+   - Verify deployment with validation queries
+5. Report deployment results
+
+**Critical**: Never deploy without dev testing first. Wrong results delivered quickly are worse than correct results delivered with a small delay.

commands/dataform-etl.md

@@ -0,0 +1,24 @@
+---
+description: Launch ETL agent for BigQuery Dataform development
+---
+
+You are launching the ETL Dataform engineer agent to handle data transformation pipeline work.
+
+**Purpose**: The ETL agent specializes in BigQuery Dataform projects, SQLX files, data quality, and pipeline development. Use this for:
+- Complex Dataform transformations
+- Pipeline troubleshooting
+- Data quality implementations
+- ELT workflow coordination
+
+**Task**: Use the Task tool with `subagent_type="etl"` to launch the ETL agent. Pass the user's request as the prompt, including:
+- What they need to accomplish
+- Any relevant context about tables, datasets, or business logic
+- Whether this is new development, modification, or troubleshooting
+
+The ETL agent has access to the dataform-engineering-fundamentals skill and will follow best practices for BigQuery Dataform development.
+
+**Example**:
+```
+User asks: "Help me create a customer metrics table"
+You launch: Task tool with subagent_type="etl" and prompt="Create a customer metrics table in Dataform following TDD workflow. Ask user about required metrics and data sources."
+```

commands/dataform-new-table.md

@@ -0,0 +1,33 @@
+---
+description: Create new Dataform table using TDD workflow
+---
+
+You are creating a new Dataform table following the Test-Driven Development (TDD) workflow from the dataform-engineering-fundamentals skill.
+
+**Workflow:**
+
+1. Invoke the dataform-engineering-fundamentals skill
+2. Ask the user about the table requirements:
+   - Table name and purpose
+   - Expected columns and their descriptions
+   - Data sources (for creating source declarations if needed)
+   - Business logic and transformations
+3. **RED Phase - Write tests first:**
+   - Create assertion file in `definitions/assertions/`
+   - Write data quality tests (duplicates, nulls, invalid values)
+   - Run tests - they should FAIL (table doesn't exist yet)
+4. **GREEN Phase - Write minimal implementation:**
+   - Create source declarations if needed
+   - Create table SQLX file with:
+     - Proper config block with type, schema
+     - Complete columns: {} documentation
+     - SQL transformation
+   - Run table creation: `dataform run --schema-suffix dev --actions <table_name>`
+   - Run tests - they should PASS
+5. **REFACTOR Phase - Improve while keeping tests passing:**
+   - Optimize query performance if needed
+   - Add partitioning/clustering if appropriate
+   - Improve documentation clarity
+6. Report completion with file locations and next steps
+
+**Critical**: Always write tests FIRST, then implementation. Tests-after means you're checking what it does, not defining what it should do.

commands/dataform-test.md

@@ -0,0 +1,18 @@
+---
+description: Test Dataform table in dev environment with safety checks
+---
+
+You are testing a Dataform table using best practices from the dataform-engineering-fundamentals skill.
+
+**Workflow:**
+
+1. Invoke the dataform-engineering-fundamentals skill
+2. Ask the user which table they want to test
+3. Follow the safety checklist:
+   - Run `dataform compile` to check syntax
+   - Run `dataform run --schema-suffix dev --dry-run --actions <table_name>` to validate SQL
+   - Run `dataform run --schema-suffix dev --actions <table_name>` to execute in dev
+   - Run basic validation queries to verify results
+4. Report results and any issues found
+
+**Critical**: Always use `--schema-suffix dev` for testing. Never test directly in production.