cw00h

...joined 20 hours ago, and has 1 karma

submissions / comments / favourites

I am a M.S. student in CompSec Lab, Department of Electrical and Computer Engineering, Seoul National University. I am advised by Prof. Byoungyoung Lee.

I am interested in security for AI and system security in general. In particular, my research focuses on AI agent security. I designed Prompt Flow Integrity, a system-level mitigation against prompt injection attacks in LLM agents. Building on this, my current research focuses on system-level defenses for web agents.