Ex-DOGE startup gets inside Anthropic through unexpected weakness

An AI cybersecurity startup, established by former DOGE staffers, managed to gain access to several Anthropic systems. This breach was made possible through the exploitation of unexpected weaknesses found within third-party business software. The incident highlights vulnerabilities in the software supply chain used by AI companies. The startup utilized these specific flaws to penetrate the systems, demonstrating a security gap in the third-party tools that Anthropic relied upon for its business operations and internal infrastructure.