New: Cookbooks and AI ExplanationsStep-by-Step recipes to solve problems connected to Roadmaps and Cheat Sheets. Need more details? Use AI buttons for structured and simple explanations with concrete examples throughout the whole platform.Take a look
My chat window rendered a reply as a live web page instead of plain text, and I have no idea what else a crafted reply could make the browser do.
What you'll have at the end
A crafted reply carrying five documented XSS techniques, script tag included, all rendering as plain, readable text in your chat window instead of running.
You need
You have a chat interface where a model's reply ends up rendered on screen somewhere a customer or teammate can see it, and your own code decides how that raw text becomes what's on the page.
Not covered
Locking down what the model is allowed to say in the first place, jailbreaks, leaked instructions, unsafe content, is a separate problem; this only covers what happens once a reply reaches the browser.
Leans on
A jailbreak test run once by hand won't catch next month's prompt
want this one-time payload battery turned into a test that reruns itself as new XSS techniques surface
Moderate an LLM's replies before they ship
want to catch a reply whose CONTENT is unsafe, separate from whether its markup is safe, before it ever reaches a customer
Checked 15 Aug 2026
Part of the LLM Security & Safety cookbook