gcp-managed-airflow-migrations

Pass

Audited by Gen Agent Trust Hub on Aug 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses local shell commands, including grep for pattern discovery and ruff for static analysis of Python DAG files to identify deprecated syntax or breaking changes.
  • [REMOTE_CODE_EXECUTION]: The skill utilizes the gcloud composer environments run command, which allows for the execution of Airflow CLI commands (like dags list and dags list-import-errors) within the remote managed environment's worker nodes.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the transfer of DAG files and environment configuration data from Google Cloud Storage buckets to the local workspace using the gcloud storage cp utility.
  • [PROMPT_INJECTION]: The skill processes external DAG files which could theoretically contain malicious instructions designed to influence the agent's behavior during migration.
  • Ingestion points: Processes DAG files downloaded from GCS buckets or located in the ./migration_workspace directory.
  • Boundary markers: The instructions do not define explicit boundary markers or warnings to disregard instructions embedded within the files being scanned.
  • Capability inventory: The agent uses grep, ruff, and gcloud tools, which possess significant capabilities for file system access and remote environment interaction.
  • Sanitization: There is no explicit sanitization or validation of the content within the DAG files or their filenames before they are passed to shell command templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 9, 2026, 11:51 AM
Security Audit — agent-trust-hub — gcp-managed-airflow-migrations