Agent security
How agents turn untrusted instructions into actions, access, and compromise.
Speaking
Talks, workshops, and live demos.
I speak about attacks on AI agents and developer tools: how we found them, how the exploit works, and what defenders can do about it. The best version usually includes a live demo.
Themes
What I speak about
From prompts to payloads
Following an attack from the prompt to code execution and infrastructure.
Software supply chains
Plugins, developer tooling, and trust decisions that quietly expand an attack surface.
Hands-on security
Labs and CTFs where participants trace the exploit instead of only hearing about it.
Selected appearances
Talks and hands-on work
Hands-on lab
From Prompts to Payloads: AtlasOps Hands-on AI/AppSec CTF
OWASP Global AppSec EU · Vienna, Austria
Talk
When Regulation Backfires: How a Vulnerable Plugin Led to an XSS Pandemic
OWASP Global AppSec EU · Barcelona, Spain
ArchiveTalk
When Regulation Backfires: How a Vulnerable Plugin Led to an XSS Pandemic
RSAC Conference · San Francisco, United States
Archive