airflow-dag-patterns
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns detected. The skill provides guidelines for Airflow DAG development, emphasizing local execution, idempotency, and data validation.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided data and files to generate Airflow DAG artifacts, creating a potential surface for indirect injection.
- Ingestion points: Inspects available files, runtime, versions, and inputs (SKILL.md).
- Boundary markers: Instructions specify to preserve user authorization boundaries and source material (SKILL.md).
- Capability inventory: Produces implementations and artifacts satisfying requested outcomes and constraints.
- Sanitization: Implements quality gates to inspect shapes, types, and units before transformation.
Audit Metadata