1 paper
Ira Ceka, Arnav Garg, Ryan Lutz +4
LLM agents are increasingly used to automate vulnerability detection, but reliable detection is challenging because agents must distinguish subtle, security-relevant fixes from ben…