AI Security & Next-Gen Terraform Ops

​Join us for our next Cloud Native Seattle Meetup at the F5 Tower in downtown Seattle! * ​5:00pm - Doors Open * ​5:30pm - Networking (pizza & drinks provided) * ​6:00pm - Scott McAllister - Why Your WAF Can’t Read a Prompt Injection * ​6:30pm - Cody Gagnon - Unlocking Terraform Day-2 Ops: From AzAPI Actions to AI-Assisted Discovery * ​7:00pm - Games and more networking **​Talk 1: Why Your WAF Can’t Read a Prompt Injection** ​Your WAF is very good at its job. It inspects the structure of every request and blocks the malformed ones. SQL injection has a shape. XSS has a shape. But a prompt injection has none: it's a plain, well-formed natural-language sentence that happens to also be an instruction. There's nothing for a signature to match. As we wire LLMs into our stacks, we've grown a new layer that inspects meaning, not structure — and none of our existing tooling watches it. This talk shows, with a live demo, how an ordinary-looking message walks past request inspection and makes an AI app leak data it was told to protect. We'll look at where inspection has to move to catch this, why indirect injection makes it worse, and what a solid defense actually looks like. ​This talk will be presented by **[Scott McAllister](https://www.linkedin.com/in/stmcallister/?utm_source=luma)**. He is a Community Evangelist at F5. He has been building software in several industries for over a decade.