The Attack and Defense Landscape of Agentic AI

Capture Summary

Systematic survey of AI agent security that analyzes design space, attack landscape, and defense mechanisms. Search result notes case studies and open challenges in securing agentic AI systems.

Relevance

• Recent broad source for attack/defense taxonomy.
• Useful for comparing academic attack classes with industry incident patterns.
• Candidate source for challenge synthesis around agent permissions, tools, memory, and autonomy.

Collection Notes

Collected as a latest research survey for Security for AI.