tools-only dg scaffold - Create Definitions
Scaffold Dagster definitions including Python objects (assets, schedules, sensors) and integration components.
ALWAYS use dg scaffold for new definitions. Do not manually create definition files.
IMPORTANT: All paths are relative to the defs/ directory, NOT the current working directory or root of the project.
Recommended Directory Structure
defs/
├── assets/
│ ├── customers.py
│ └── sales/
│ └── revenue.py
├── schedules/
│ └── daily_refresh.py
└── sensors/
└── file_watcher.py
Python Objects
dg scaffold defs dagster.asset <path>.py
dg scaffold defs dagster.schedule <path>.py
dg scaffold defs dagster.sensor <path>.py
Examples:
# Asset
dg scaffold defs dagster.asset assets/customers.py
dg scaffold defs dagster.asset assets/sales/revenue.py
# Schedule
dg scaffold defs dagster.schedule schedules/daily_refresh.py
# Sensor
dg scaffold defs dagster.sensor sensors/file_watcher.py
Integration Components
dg scaffold defs <component_type> <name> [--json-params '{...}']
Examples:
# dbt
dg scaffold defs dagster_dbt.DbtProjectComponent my_dbt --json-params '{
"project_dir": "dbt_project"
}'
# Sling
dg scaffold defs dagster_sling.SlingReplicationComponent my_sling --json-params '{
"replication_config": "replication.yaml"
}'
Find available components:
dg list components
dg list components --package dagster_dbt
For integration-specific configuration and parameters, use the dagster-integrations skill.
Custom Components
# Reusable component type
dg scaffold component MyComponentType
Creates a new component class that can be instantiated multiple times via YAML.
Validation Checklist
After scaffolding, always run:
dg check defs && dg list defs
This verifies:
- Definition syntax is valid
- All imports resolve correctly
- Definition appears in the registry
Generated Structure
Python objects: Single .py file at specified path.
Integration components: Directory with defs.yaml:
defs/
└── my_dbt/
└── defs.yaml