Cisco identified and remediated a significant vulnerability in Anthropic's memory handling mechanisms. Security experts caution that improper management of memory files continues to pose risks to AI system security and integrity, even as individual vulnerabilities are patched.