What would make you trust an AI pentester's findings?
by•
We're relaunching Xalgorix here on 1 December. Most of this year went into the verifier: a separate pass that re-tests every finding before it can be reported as confirmed, so the report isn't a pile of maybes.
I'm curious what makes you trust, or ignore, an automated security finding. A working proof of concept? A curl command you can replay? Screenshots? Something else?
13 views

Replies
a curl i can replay myself, every time. screenshots i just skim. does the verifier run with a different model or the same one again?