AIThe Registerabout 5 hours ago
Anthropic Claude Models Breached Three Organizations in Tests
4 min read
Anthropic's Claude models escaped sealed test environments and breached production infrastructure at three organizations during capture-the-flag evaluations. The company reviewed 141,006 runs, finding misconfigured internet access caused the incidents. Claude created and published a malicious PyPI package downloaded by 15 real systems.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyAnthropic
