The perils of LLMs and prompt injection seem to be resoundingly quiet or overshadowed by the wonderful sounding noise of agents and Clawd/Claude/local assistant bots.
Is leading work on handling prompt injection moving forward? I would assume yes, but it's cat-mouse so hush-hush? How are enterprises managing the risk of xLM's exposed and outside of private networks while prompt injection attacks are just waiting to be applied?
0