deepstream-dev

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime path can ingest outsider-authored free text via the dynamic multi-stream REST API exposed by nvmultiurisrcbin (e.g., POST /api/v1/stream/add supplies source URIs that the pipeline can then fetch/consume through nvurisrcbin), meaning an outsider can submit poison text into the feed the workflow reads.