Former DOGE Staffers Say AI Agents Breached Anthropic in 12 Hours
WHO:
The Washington Post - Cathedral is a months-old cybersecurity start-up founded by former Department of Government Efficiency (DOGE) staffers.
The Washington Post - Anthropic makes the Claude chatbot.
Techmeme - Cathedral is a months-old cybersecurity startup founded by former DOGE staffers; Anthropic makes the Claude chatbot.
WHAT:
Inc - Anthropic reported in its S-1 filing that it had received applications from people with “ulterior” motives.
The Washington Post - Cathedral told Anthropic that its researchers had penetrated sensitive internal data and messaging systems at the artificial intelligence company.
Techmeme - Cathedral said it breached Anthropic’s internal systems using AI agents.
Techmeme - The Washington Post reported that Cathedral’s researchers were able to penetrate sensitive systems quickly.
WHEN:
The Washington Post - Cathedral’s reported breach took 12 hours.
Techmeme - Cathedral said it breached Anthropic’s systems in under 12 hours.
The Washington Post - Cathedral is described as a “months-old” start-up.
WHY:
Inc - Anthropic described applicants with “ulterior” motives as an “insider threat.”
HOW:
The Washington Post - The account of the penetration appeared in a draft report by Cathedral reviewed by The Washington Post.
Techmeme - Cathedral said its researchers used AI agents to penetrate Anthropic’s internal systems.
(UPDATED October 26, 2026 19:23 UTC)
Cathedral, a cybersecurity startup founded by former DOGE staffers, says its researchers used AI agents to breach sensitive internal systems at Anthropic, the company behind the Claude chatbot. Cathedral said the breach took under 12 hours, and its account appeared in a draft report. Anthropic has said it received applications from people with “ulterior” motives, which it described as an insider threat.
--
(UPDATED October 26, 2026 14:49 UTC)
Cathedral, a cybersecurity startup founded by former Department of Government Efficiency staffers, said its researchers used AI agents to breach sensitive internal systems at Anthropic, the company behind the Claude chatbot. Cathedral said the researchers got in in under 12 hours. The account appeared in a draft report reviewed by The Washington Post.