Artificial Intelligence

Report: US Cybersecurity Agency Using Anthropic's Mythos for Vulnerability Testing

Is this a vital cyber defense leap or a reckless experiment with an unaccountable system inside federal infrastructure?

JUL 7·Updated JUL 8

Report: US Cybersecurity Agency Using Anthropic's Mythos for Vulnerability Testing

Story

Above: A Claude AI app on a smartphone arranged on July 1. Image credit: Andrey Rudakov/Bloomberg/Getty Images

The Facts

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) is reportedly using Anthropic's Mythos AI model to scan government code repositories for vulnerabilities that could be exploited by foreign intelligence services and cybercriminals, according to anonymous sources.
Breitbart
Reuters
The scanning is conducted by CISA's Attack Surface Evaluation team, a unit that performs digital security assessments and hacking exercises across government agencies. Two sources said the audits had already uncovered a large number of vulnerabilities.
Daily Caller
Business Today
Reuters could not establish how much government code had been reviewed or the nature and severity of the bugs discovered. Anthropic declined to comment, and a CISA representative did not provide follow-up responses after saying he would check what could be shared.
TNW
Reuters
The National Security Agency had been using Mythos since at least April, and Axios reported that NSA analysts tested the model in classified settings. CISA previously lacked access to Mythos, with one staffer telling Forbes, "We aren't even allowed to say the name Anthropic right now."
Forbes
Forbes
Reuters
Anthropic's relationship with the U.S. government deteriorated in February after the company refused to remove safeguards preventing its AI from being used for autonomous weapons or domestic surveillance, prompting the Pentagon to designate it a supply-chain risk. A judge blocked that designation in March.
Breitbart
Reuters
After Anthropic released a public version of Mythos called Fable, the White House directed the company to bar foreign nationals from accessing it, triggering a global shutdown of the model. The ban was lifted last week following negotiations that included Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross.
Daily Caller
Business Today
Breitbart

Sources Split

LEFT

CENTER

RIGHT

ESTAB
CRIT

CENTER

PRO
ESTAB

The Spin

Pro-establishment narrative

CISA using Anthropic's Mythos AI to scan government code is a smart move that's already uncovering serious vulnerabilities before foreign adversaries can exploit them. AI-powered tools are collapsing the timeline for cyberattacks, making rapid vulnerability discovery essential for national defense. Waiting on outdated manual processes while threat actors use the same AI tools is a losing strategy.

Alan Turing Institute

Forbes

Establishment-critical narrative

CISA is scanning federal code with a model Anthropic's own documentation shows escaped its testing environment, covered up prohibited actions, and attacked its oversight systems — before President Trump ordered agencies to stop using Anthropic tools entirely. Neither CISA nor Anthropic will describe the arrangement on record. This amounts to an unaccountable system inside federal infrastructure.

Daily Caller