article-analyzer

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

• Third-party content exposure detected (high risk: 0.80). The skill both accepts user-provided articles and explicitly requires performing "联网检索" / fact-checking and "背景补充" by searching public sources (see the "联网检索要求" and "检索类型" sections), meaning the agent will fetch and read untrusted third-party web/user-generated content as part of its workflow.