brand

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

• Third-party content exposure detected (high risk: 0.80). The skill's Phase 1 explicitly asks for "images, links, or files to share" and instructs "If they share images: Read them using the Read tool" (Phase 1 Gather References), meaning the agent will fetch and analyze arbitrary user-provided or web-hosted third‑party content that will directly influence the generated brand directions and subsequent actions.