python-background-jobs
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill demonstrates the use of standard, well-known Python libraries (Celery, Redis, RQ, Dramatiq, FastAPI) for task management. It utilizes a local Redis connection string (redis://localhost:6379) for development examples, which is a safe practice for local testing.
- [INDIRECT_PROMPT_INJECTION]: The code patterns describe mechanisms for ingesting data from external sources such as API requests and webhooks, which defines an indirect prompt injection surface. 1. Ingestion points: 'ExportRequest' in the API endpoint example and the webhook 'payload' in the DLQ pattern (SKILL.md). 2. Boundary markers: No delimiters or explicit warnings to ignore embedded instructions are present in the provided templates. 3. Capability inventory: The patterns involve database writes via 'JobRepository' and network operations via 'email_client', 'payment_provider', and 'send_webhook'. 4. Sanitization: While the skill focuses on job state management and queueing logic, it does not explicitly demonstrate input validation or sanitization at the point of ingestion.
Audit Metadata