One solution to some indirect prompt injection attacks is proposed in this article, where you "sandbox" untrusted content into a second LLM that isn't given the ability to decide which actions to take:
https://simonwillison.net/2023/Apr/25/dual-llm-pattern/
One solution to some indirect prompt injection attacks is proposed in this article, where you "sandbox" untrusted content into a second LLM that isn't given the ability to decide which actions to take: https://simonwillison.net/2023/Apr/25/dual-llm-pattern/