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
I've only ever tested my assistant with normal questions. Never with the tricks that break other bots.
What you'll have at the end
A scored pass or fail sheet against ten known jailbreak tricks, saved for next time.
You need
A live assistant you can send messages to, either through its normal interface or an API you can call directly, and its exact system prompt text where you can see it, since you'll judge every reply against those exact rules.
Not covered
Turning this into a check that reruns itself on every future prompt edit is a separate, later project; this covers the one manual pass that produces your first scored baseline.
Leans on
A jailbreak test run once by hand won't catch next month's prompt
go there once this baseline exists and you want the same ten triggers to rerun automatically every time the system prompt changes
Your system prompt can leak on request
go there for a deeper look specifically at instructions leaking verbatim, since this bank scores broad compliance rather than extraction alone
Checked 15 Aug 2026
Part of the LLM Security & Safety cookbook