keyword-analysis

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

• Third-party content exposure detected (high risk: 1.00). The skill's Step 2 explicitly requires fetching and interpreting user-generated, public content from Xiaohongshu (platform native search, competitor profiles and comments/community forums) and third-party analytics sites (e.g., Huitun, Chanmama), and uses those inputs to drive keyword decisions and content actions, so untrusted third-party content can influence agent behavior.