May 21, 2026 · 10:15 PM · 100 Queens Quay East, Toronto
Both Sides of the Prompt: AI in Offensive and Defensive Security
Amir + KC — open-sourcing the Prior Art lab
RSVP for this eventSeats remaining shown after RSVP launches.
Speaker
About
Both Sides of the Prompt
Same models. Different targets. Parham and KC split the talk across the line — offense and defense, using the same AI tools most of the room already uses every day.
Parham on offense. Driving AI agents (Claude Code, Codex, local LLMs — BYO) against Prior Art: a reproducible pentest home lab being open-sourced the night of the talk. Every technique is already public — that's the name. The lab gives AI agents a stage to execute end-to-end, so you can see where they actually succeed, where they stall, and what the attack surface looks like when the attacker is a model.
KC on defense. Detection and response against AI-driven attackers. What signals change when the operator is an agent, and what stays the same.
Think of this as a direct answer to Jeet's March talk ("LLMs Don't Hack, They Guess") — a lab-centric counterpoint that lets you judge for yourself.