multi-patent-avoidance
Warn
Audited by Snyk on Aug 27, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (medium risk: 0.30). 该技能在步骤2-8中直接读取“客户提供的风险专利清单/各专利的独立权利要求文本(PDF/谷歌专利/Patsnap)”并对其进行提取、矩阵化与逐格扫描,属于用户/外部提交的自由文本被模型在运行时直接摄取(通过专利权利要求文本内容)。
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata