AI startup claims it hacked Anthropic in 12 hours, gaining access to employee details and internal messages
follow us
Add us as a preferred source on Google
- AI startup Cathedral used autonomous agents to breach Anthropic systems via weaknesses in third-party software
- Agents accessed human resources, sales, and IT data, exposing limited customer and partner information.
- The incident highlights how AI-powered attackers can quickly exploit suppliers’ weaknesses despite strong internal defenses.
An AI startup apparently gave Anthropic a taste of its own medicine when it used AI agents to break into the company and expose sensitive data.
The company in question is called Cathedral, which was founded by former members of the defunct US Department of Government Efficiency (DOGE), a US government initiative established to reduce federal spending and bureaucracy, while modernizing government technology and operations.
Anthropic is running a dual bug bounty program through HackerOne. One part focuses on finding technical flaws in Anthropic’s systems, while the other is an invitation-only program that rewards researchers for finding universal leaks that bypass Claude’s security protections.
Latest videos ofTechnologyRadar
Cathedral apparently shared its results with Anthropic through the bug bounty program. After their human investigators tried unsuccessfully to break in for a few days, they deployed AI agents that took less than 12 hours to find a way in, exploiting flaws in third-party enterprise software.
violating the rules
After accessing the company network, agents accessed meeting transcripts, salary information and job offers, as well as Slack messages. Apparently, the employees’ conversations with Claude were not exposed.
Anthropic confirmed the news, saying Cathedral accessed a “limited set” of sales, human resources and IT systems, and compromised contact information and basic user account records belonging to certain customers and partners. The source code and weights of the AI model were not accessed.
“Our monitoring detected some of the Cathedral activity before it was reported to us,” Anthropic said in a statement shared with Washington Post. “We closed access to them the same day.”
According GizmodoAnthropic accused Cathedral of violating program rules by failing to disclose the full extent of its activity. We don’t know if that means Anthropic won’t release the reward payment, but it sure seems that way.
However, paying the reward could be the least of Anthropic’s worries. The company is currently preparing for its Initial Public Offering (IPO), with a valuation of $1.4 billion. Leaking sensitive data to anyone, including white hat hackers, is definitely not a good sign.
A ‘group of children’ wreaking havoc
“The fact that weaknesses in ordinary enterprise software can compromise the infrastructure of even the most advanced organizations should be a wake-up call for every business in America,” he said. Washington Post cited the Cathedral report.
At the same time, a senior Trump administration official allegedly told him Through the independentthat “if a bunch of kids who just started a company can do this, then potentially the Chinese are not far behind.”
Cathedral can hardly be described as “a bunch of kids”: it’s an AI startup focused on offensive and defensive cyber operations for the US military and government. Its founders include Gavin Kliger (former Pentagon chief data officer), Luke Farritor (best known for helping decipher ancient Roman scrolls using machine learning), Marko Elez (former SpaceX employee), and Jack Stein (apparently worked as an investment banking associate at Salem Partners before joining DOGE).
The cybersecurity community had something to say about it. Jacob Krell, senior director of secure AI and cybersecurity solutions at Suzu Labs, was not particularly gradual, saying that if third-party vendors are a risk everywhere, then they are a risk at Anthropic.
“Cathedral agents entered Anthropic the same way attackers enter everyone else, through vendors,” Krell said. “Security maturity is limited by the weakest vendor you integrate with, and no amount of internal engineering changes that.”
Darin Fredde, senior director of technical marketing engineering at Ridge Security, praised Anthropic’s response: “Anthropic deserves credit for reporting that its monitoring detected some of the activity and for taking containment measures. But the real remediation measure is independent verification that those attack paths are closed. We need to stop confusing a successful patch with a verified security result. Otherwise, we’re just locking the front door and hoping no one discovers the spare key.”
The best antivirus for all budgets
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds.
Post Comment