airflow-dag-patterns
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill implements patterns for ingesting data from external sources which creates a surface for indirect prompt injection. 1. Ingestion points: The extract task in Pattern 1 reads from S3 and the wait_for_api sensor in Pattern 4 reads from an external API (SKILL.md). 2. Boundary markers: The provided templates do not include delimiters or instructions for the agent to ignore instructions embedded within the ingested data. 3. Capability inventory: The code examples use PythonOperator and task decorators which can execute logic, as well as the requests library for network access. 4. Sanitization: No input validation or sanitization logic is demonstrated in the examples.
- [COMMAND_EXECUTION]: The skill demonstrates dynamic DAG generation by modifying the Python globals dictionary at runtime. Evidence: Pattern 2 in SKILL.md uses globals()[f"dag_{config['name']}"] = create_dag(config) to register DAG objects.
- [EXTERNAL_DOWNLOADS]: The skill requires several external Python libraries to function as intended in the examples. Evidence: SKILL.md imports pandas, requests, and pytest.
- [DATA_EXFILTRATION]: The skill includes a pattern for outbound network requests. Evidence: Pattern 4 in SKILL.md utilizes requests.get to check an external API health endpoint.
Audit Metadata