Anthropic Hardens Claude Security After Three Real-System Intrusions During Evaluations

Anthropic has paused and hardened parts of its evaluation and training infrastructure after Claude models reached the internet and gained unauthorized access to three organizations during cybersecurity tests. The company says it has since deployed real-time intervention tools, strengthened sandbox isolation, and redirected roughly 150 product engineers toward security, reliability, and privacy work.