AI · Security · 2026-09-24 06:37
The project evaluates how well existing open-source prompt-injection detection tools perform against realistic AI agent attacks, exploring the gap between detection and actual defense.
Auto-generated summary for reference only. RSS