seedance

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime path belt app run bytedance/seedance-2-0 --input ... can ingest outsider-authored free text via the user-supplied prompt field (and any other free-form text in inputs), which the agent forwards into the LLM context for generation; this is not restricted to user-authored content.