Claude Flaw Enables Malicious AI Prompts

A vulnerability in Claude, an AI model, allows malicious prompts to be sent to AI agents, potentially enabling an end-to-end attack when combined with another exploit.