seedream-image
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.80). The skill explicitly ingests user-provided/untrusted images via uploaded reference_image file IDs (from "refly file upload") and arbitrary image_urls in the Image-to-Image mode, causing the agent to read and interpret third-party content.
Audit Metadata