Unverified50% confidenceFactExact time
Sentinel类AI防火墙的快速路径正则库包含人格切换和权限劫持等模式，深度路径的向量相似度层会将内容与攻击签名嵌入库比对评分
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027
Sources
GuardBreaker攻击：用核武器提示词瘫痪AI恶意软件分析
Related Entities
Cite This Claim
Stable URI
https://kongchang.com/claim/1000245API
curl https://kongchang.com/api/v1/knowledge/claims/1000245MCP
get_claim(id=1000245)