[danieltk76]

At Vulnetic ai we deal with this by having a separate validator agent attack the vulnerability reported using a novel perspective separate of the hacking agent.