story-short-scan

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

• Third-party content exposure detected (high risk: 1.00). The skill's Phase 1.5 "确定数据来源" explicitly instructs using WebSearch/WebFetch and an agent-browser CDP to fetch public platform pages (e.g., 知乎盐言、番茄短篇、小红书) or user-provided links and then uses that content in Phase 2 analysis and report generation, exposing the agent to untrusted user-generated third-party content that can influence decisions.