Automating Agent Hijacking via Structural Template Injection
arXiv:2602.16958v1 Announce Type: new Abstract: Agent hijacking, highlighted by OWASP as a critical threat to the Large Language Model (LLM) ecosystem, enables adversaries to manipulate …
Xinhao Deng, Jiaqing Wu, Miao Chen, Yue Xiao, Ke Xu, Qi Li
6 views
