Redpanda reports on a flaw called CoreBreak in major AI agent platforms, exposing inadequate guardrails.
From
To