In a new revelation. the US cyber security unit namely CISA (Cybersecurity and Infrastructure Security Agency) is employing Anthropic’s Mythos to go through government code repositories in order to identify and fix security vulnerabilities. Mythos is a frontier AI model capable of detecting, fixing as well as exploiting vulnerabilities far more quickly compared to older AI models. In other words, it is a potent tool that can be utilized for effective cyber defense as well as mounting massive cyberattacks. Anthropic is a leading company in the space of state-of-the-art AI technology.

Strained Relationship between the US Government and Anthropic

For your information, the US government and Anthropic have been at loggerheads for some time now. The reason being the AI giant refused to let its Mythos application be used by the US government for building autonomous weapons as well as for carrying out domestic surveillance. In a response to the refusal, the US government blacklisted the AI company. Usually, the designation is given to entities carrying out espionage. In March, the blacklisting was turned down by a court.

Recognition by Government Agencies of Mythos’s Impressive Capabilities

The Cybersecurity and Infrastructure Security Agency is employing Mythos to go through government code repositories to identify bugs that could be leveraged by foreign intelligence agencies as well as cybercriminals. For your information NSA, the National Security Agency was utilizing Mythos since April in spite of the government blacklisting. As per the New York Times, NSA analysts found the Frontier AI technology quite impressive.

Secrecy Regarding the Exact Volume and Type of Vulnerabilities

Both Anthropic and CISA are mum about the significant developments. The examination of code belonging to the US government is being executed by CISA’s Attack Surface Evaluation group. The latter is responsible for carrying out digital security evaluation as well as simulated attacks on federal websites. However, full details of the sensitive results have not been disclosed. When the public version of Mythos, namely Fable was launched, the US government asked Anthropic not to give access to foreign nationals. However, the ban was taken off the previous week.

FAQs

1. What is CISA using Anthropic Mythos for?
CISA is using Anthropic’s Mythos AI model to analyze U.S. government code repositories for security vulnerabilities. The goal is to identify and help fix weaknesses before they can be exploited by hackers or foreign adversaries.

2. Why is AI being used to detect software vulnerabilities?
Advanced AI models can review large amounts of code much faster than humans. They can identify potential security flaws, suggest fixes, and help cybersecurity teams improve software security more efficiently.

3. What are government code repositories?
Government code repositories are secure storage locations where software code used by federal agencies is kept. These repositories contain applications and systems that support government operations and need strong protection from cyber threats.

4. Can AI like Mythos also be misused?
Yes. While AI can strengthen cybersecurity by finding and fixing vulnerabilities, similar technology could also be used by malicious actors to discover weaknesses in software. This is why access to advanced AI models is often carefully controlled.

5. Why is this development important for cybersecurity?
The use of AI by CISA shows how governments are adopting advanced technology to improve cyber defenses. Faster vulnerability detection can help prevent cyberattacks, protect sensitive government systems, and enhance national cybersecurity.

SOURCES:-

https://www.ndtv.com/world-news/us-cybersecurity-agency-uses-anthropic-ai-model-mythos-to-audit-government-software-11736532

https://www.firstpost.com/world/us-cybersecurity-agency-cisa-using-anthropics-mythos-to-audit-govt-software-report-14029438.html

https://www.wionews.com/world/us-cybersecurity-agency-is-reportedly-using-anthropic-s-mythos-for-auditing-govt-code-what-does-it-mean-1783414731952

https://www.businesstoday.in/technology/news/story/us-cyber-agency-cisa-uses-anthropics-mythos-to-audit-government-software-541353-2026-07-07

https://www.securityweek.com/cisa-reportedly-using-anthropics-mythos-to-scan-government-software-for-flaws/

https://www.ndtvprofit.com/business/us-government-turns-to-anthropics-mythos-ai-for-cybersecurity-audits-report-11736607

https://www.devdiscourse.com/article/technology/3946113-mythos-a-new-era-in-cyber-defense-amidst-government-struggles

https://www.technology.org/2026/07/07/cisa-anthropic-mythos-audit-government-code/

https://cyberpress.org/anthropic-mythos-ai-driven-vulnerability/

https://www.msn.com/en-us/news/technology/us-cyber-agency-is-using-anthropic-s-mythos-to-audit-government-code-reuters/ar-AA27mgOp?apiversion=v2&domshim=1&noservercache=1&noservertelemetry=1&batchservertelemetry=1&renderwebcomponents=1&wcseo=1