What problem does it solve? Writing Airflow DAGs for Managed Service for Apache Airflow (formerly Cloud Composer) requires matching the target environment's Airflow version and installed packages, and undeployed DAGs often fail with import errors only discovered after deployment. This Skill guides environment discovery, version-aware authoring, and local and remote validation so DAGs parse correctly before and after deployment. ## Core Features & Use Cases - Environment Context Discovery: Uses gcloud commands to inspect the target environment's Airflow image version, installed PyPI packages, and DAGs GCS bucket before writing code. - Best-Practice Authoring: Enforces idempotent tasks, no top-level code execution, explicit catchup settings, and Airflow Variables/Connections instead of hardcoded credentials, with Airflow 2 vs 3 compatibility guidance. - Local and Remote Validation: Runs ruff linting, composer-dev import-error checks, GCS deployment, and Airflow CLI verification against the target environment. - Use Case: When asked to create a new DAG for a Cloud Composer environment, the Skill first discovers the environment's Airflow version, writes compatible DAG code, then deploys it to the GCS bucket and confirms there are no import errors. ## Quick Start Ask the AI to create and validate a new Airflow DAG for your Managed Airflow environment, providing the environment name and GCP region.
