Prompt file imported from asabocinski/bravo-edd (
.codex/prompts/pipeline.md). Copyright stays with the author.
Pipeline Command
Scaffold a data pipeline (Airflow, Dagster) with best-practice patterns
Usage
/pipeline <description-or-file>
Examples
/pipeline "Daily orders ETL from Postgres to Snowflake"
/pipeline "Kafka → staging → dbt → marts with hourly refresh"
/pipeline requirements/pipeline-spec.md
What This Command Does
- Invokes the pipeline-architect agent
- Analyzes your description or requirements file
- Loads KB patterns from
airflowanddbtdomains - Generates:
- DAG structure (Airflow or Dagster)
- Task definitions with dependencies
- Error handling and retry configuration
- Sensor/trigger patterns for scheduling
Agent Delegation
| Agent | Role |
|---|---|
pipeline-architect |
Primary — DAG design, task orchestration |
spark-engineer |
Escalation — when pipeline includes Spark jobs |
dbt-specialist |
Escalation — when pipeline includes dbt models |
KB Domains Used
airflow— DAG patterns, operators, sensorsdbt— model execution, incremental strategiesdata-quality— quality gates between pipeline stages
Output
The agent generates pipeline code files and a summary of the DAG structure with task dependencies.
