jacky-opc

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

• Third-party content exposure detected (high risk: 1.00). The skill's runtime workflow explicitly mandates proactive "联网搜索" (Step 3 in SKILL.md) and several modules state "联网补充信息"/"联网更新", while references/路径指南.md's "搜索顺序" includes community sources (Reddit, Hacker News, X), so the agent will fetch and read untrusted public/user-generated web content that can materially influence decisions and subsequent actions.