live-promotion

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to read and respond to live Xiaohongshu viewer comments and user-generated testimonials during streams (e.g., "Read comments continuously", "Respond to questions", "Feature user testimonials (read from comments)"), which are untrusted third-party inputs that can directly influence what the agent says or does.