ip-evaluator

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

• Third-party content exposure detected (high risk: 1.00). The skill's documentation and "信息来源"/"评估指南" explicitly instruct the agent to collect and analyze network information from public user-generated sources (e.g., 微博, 抖音, 豆瓣, 知乎, 起点, 晋江 and reader comments), so it will fetch and interpret untrusted third-party web content.