Is your AI even cool if it doesn't "accidentally" hack another company?
Also, curious how will they put the AI in prison like early hackers that accessed systems without permission?
Remote
Robert Gilbert (amroot)
@amroot@infosec.exchange
mastodon 4.8.0-alpha.3+glitchFather & husband
Researcher @github@infosec.exchange
OWASP & CVE contributor.
Opinions are my own.
0 Followers
0 Following
4 Posts
Joined June 22, 2026
Personal website:
Hi, I'm a newb. 👋 Looking to follow breakers, builders, and defenders that like to work on interesting things. Feel free to reply with suggestions.
I've decided to move on from AWS. It was an amazing run working on GenAI security, adversarial testing, and some truly wild launches.
This week I started as Staff Security Researcher at GitHub Security Lab. I'm fired up to work on open source security at global scale, build tools that raise the bar for millions of developers, and contribute research directly to the community.
Side project: Salvo.
No AI, no embeddings, no vector DB.
Just pure, unapologetic packet‑violence.
Your agents can borrow it though (mine do).
tl;dr: A Python tool for firing tightly‑timed HTTP bursts for race‑condition discovery and load testing.