skills/googlecloudplatform/data-agent-kit-plugin/gcp-managed-airflow-dag-authoring/Gen Agent Trust Hub
gcp-managed-airflow-dag-authoring
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions to use official CLI tools like
gcloudandcomposer-devfor environment discovery and DAG validation. These commands are standard for managing Managed Service for Apache Airflow (Cloud Composer) and do not involve suspicious remote execution. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and validate user-authored Python files (Airflow DAGs).
- Ingestion points: The skill interacts with local DAG files specified by the user in
SKILL.md(e.g.,path/to/dag.py). - Boundary markers: None are explicitly defined, as the skill operates on the code itself.
- Capability inventory: The skill uses tools for static analysis (
ruff,pylint), file operations (cp), and cloud deployment (gcloud storage cp,gcloud composer environments run). - Sanitization: Standard for development workflows; the skill instructs the agent to perform linting and parsing checks before deployment.
- [SAFE]: The guide includes security-positive instructions, specifically advising users to never hardcode credentials and instead use Airflow's built-in secure Variable and Connection mechanisms.
Audit Metadata