In July 2023, BSI issued a cybersecurity warning regarding indirect prompt injections, which are an intrinsic vulnerability in application-integrated Large Language Models. This documents provides an English translation of the cybersecurity warning."

Version 1.0