Former DOGE Staffers' Startup Claims AI Agents Accessed Sensitive Anthropic Data

A cybersecurity startup founded by former US DOGE staffers allegedly accessed sensitive Anthropic internal data using AI agents, according to a draft report reviewed by The Washington Post. The breach reportedly exposed meeting transcripts, salary details, and Slack messages. Cathedral, a months-old startup, exploited vulnerabilities in third-party business software used for administrative and sales tasks, running AI agents for under 12 hours to reach the data. It disclosed findings to Anthropic and software vendors this month, with Anthropic acknowledging the breach of limited sales, HR, and IT systems. Anthropic disabled access pathways the same day, confirming exposure of business contact details and basic user records, but not customer messages, AI code, or model weights. The incident follows growing cybersecurity concerns, including Claude's recent access to Georgia vote counts.