Cathedral Researchers Reached Anthropic Internal Systems in 12 Hours

Cybersecurity start-up Cathedral told Anthropic that its researchers got into sensitive internal systems at the company after running AI agents for less than 12 hours, The Washington Post reported, citing a draft Cathedral report.

The researchers exploited vulnerabilities in business software that Anthropic and many other companies use for administrative and sales work. According to the report, they found paths to meeting transcripts, salary and job offer details, and Slack messages about Anthropic's advanced models. Cathedral probed Anthropic under the company's program that invites outside security researchers to test its systems, and disclosed the findings earlier this month.

Anthropic confirmed that Cathedral reached "a limited set" of its sales, human resources and IT systems. It said it shut off the access paths and briefed federal officials. Cathedral was founded by former members of the U.S. DOGE Service, sells AI-powered security tools to government and corporate clients, and has been valued by investors at $1.4 billion.