Review
Open access
2026
A survey on AI agent security: Reasoning, Acting, and Self-Evolving
Artificial intelligence (AI) agents are rapidly evolving into autonomous systems capable of reasoning, acting, and continual evolving. Their increasing autonomy enables powerful real-world applications but also introduces security risks throughout the entire agent execution cycle. However, the security challenges arisi...
Dai-Zong Liu, Tian-Yao Luo, Shu-Wei Huang et al.
· AI Plus · 0 citations