gcp-managed-airflow-migrations
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
gcloudCLI and standard shell utilities such asgrepandruffto inspect Cloud Composer environments, list active DAGs, and perform static analysis on code files. These operations are restricted to the migration workspace and the user's authorized GCP environment. - [EXTERNAL_DOWNLOADS]: Instructions are provided for downloading DAG files and configuration from Google Cloud Storage (GCS) buckets using
gcloud storage cp. These downloads target the user's own cloud infrastructure and are only performed upon explicit request. - [INDIRECT_PROMPT_INJECTION]: The skill processes external code by downloading and scanning DAG source files, creating an attack surface for indirect prompt injection.
- Ingestion points: Source code files downloaded from GCS buckets into the
./migration_workspace/dagsdirectory. - Boundary markers: None explicitly defined for the scanned content.
- Capability inventory: The skill can execute shell commands (
grep,ruff,gcloud), perform local file system operations, and upload modified code back to Cloud Storage. - Sanitization: No specific sanitization or filtering of the DAG file content is performed prior to the agent's analysis or remediation steps.
Audit Metadata