Logic-Layer Prompt Control Injection (LPCI): A Novel Security Vulnerability Class in Agentic Systems
Explores LPCI, a new security vulnerability in agentic AI, its lifecycle, attack methods, and proposed defenses.
Abstract: Automation plays an important and irreplaceable part in many modern industries, where automated control systems gather, analyze, and respond to information almost autonomously to increase ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results