Defending the AI Attack Surface in the Claude Mythos Era.
You don't need another summary of what Mythos did.
You need to see what that same capability finds when it's pointed at the open internet — at your kind of infrastructure.
So that's what we're going to do. Live. On the call.
CloudSEK's Research Lead will open with what actually changed on April 7 — then stop talking and start scanning.
The scan is the session.
Using AIVigil, live, in real time, the team recreates the exact discovery pattern Mythos demonstrated — aimed at the new frontier of exposure that traditional monitoring wasn't built to see:
🔓 Exposed MCP servers
The connective tissue of agentic AI. Spinning up faster than security teams can inventory them — increasingly public, increasingly unguarded, and almost never on anyone's asset list.
🔑 Leaked AI keys
Credentials to your models, your pipelines, your agents — sitting in public repos and paste sites right now. One key is one door. Most orgs don't know how many they've left open.
🤖 Agentic workflows in the wild
Autonomous systems making decisions and taking actions with no human watching. Powerful when you run them. Dangerous when someone else finds them first.
Built for the people who'll be asked to explain this
CISOs and security leaders. Heads of AppSec and threat intel. Anyone whose board has already forwarded them a Mythos headline with three words: "Are we exposed?"
If you'd rather answer that question with a live scan than a guess, this is your room.
Zoom link will be shared post-registration.