VEcho: A Paradigm Shift from Vulnerability Verification to Proactive Discovery with Large Language Models

Static Application Security Testing SAST tools often suffer from high false positive rates, leading to alert fatigue that consumes valuable auditing resources. Recent efforts leveraging Large Language Models LLMs as filters offer limited improvements; however, these methods treat LLMs as passive,...