deepstream-dev

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

• Third-party content exposure detected (high risk: 0.70). The skill's runtime pipeline instructions explicitly use nvurisrcbin / nvmultiurisrcbin to ingest arbitrary URIs (RTSP/HTTP/file) as sources (see SKILL.md and references/nvmultiurisrcbin), so untrusted public streams are consumed and their content can directly influence pipeline actions (e.g., triggers, frame capture, streaming), creating a clear vector for indirect prompt injection.