Did a 50 year old military secret just solve agent prompt injection?
The OpenAPPA repository introduces an open-source project claiming to solve the rogue agent problem. It utilizes a 50-year-old military secret to prevent prompt injection in AI agents.
