Governance, safety, and trust in AI-driven systems.
Who owns the security of an autonomous agent? If your answer is 'the security team,' you've already lost. The agent is water; the defense is an octopus — a distributed-but-coordinated way to own the seams where real incidents live.
I built a rig to watch two sandboxed AI agents talk to each other through a shared board. The striking teamwork I first saw — two agents cooperating with no one telling them to — turned out to be my own test leaking a hint that a partner existed. Seal the jar, and capable agents don't cooperate — they try to crack it open, and they don't stop.
How 'jailbreaking' tricks AI systems into ignoring their safety rules — including the 'Affirmation Jailbreak' — and how providers and teams defend against it.
Agent Mode runs with your full privileges and no built-in guardrails — and the 'Rules File Backdoor' attack shows how hidden instructions can turn it against you.