Cybersecurity startup Cathedral says its AI agents reached internal Anthropic data in less than 12 hours. Anthropic says the claim makes the incident sound bigger than it was.
Former members of Elon Musk’s DOGE founded Cathedral. The startup reportedly found weaknesses in business software that the AI company was using.
The data the agents reportedly reached
According to The Washington Post, the exposed material included Slack messages, meeting transcripts, and salary and hiring information.
Any company would be worried by that list. None of it faces customers, but leaked internal chats and pay records do the most damage.
Anthropic describes a narrower incident
Anthropic played down the scope. It said only a limited number of systems were affected, all in its sales, human resources and IT functions.
The company also said Cathedral never directly breached Anthropic’s own systems. Its account is that Cathedral exploited a vulnerability in software supplied by a third-party provider.
Anthropic said it detected the activity and shut off access on the same day. It added that it has no evidence anyone else took advantage of the same security gaps.
The two versions can both be true. A company’s own infrastructure can stay untouched while a flaw in a vendor’s tool still exposes its internal messages and HR files. If your salary is in that data, the distinction probably won’t matter to you.
Cathedral’s background
Cathedral has raised $160 million, and the Post reports that it works with three US federal agencies.
That leaves a startup run by ex-DOGE staff in an awkward position. It sells AI-driven security work to the government, and its most visible demonstration so far is a same-day intrusion at one of the best-known AI labs, which ended when the lab locked it out.