WIRED: An Accenture employee’s personal Claude Code agent ran on a Pentagon network for two months

The allegation

WIRED published a report on Oct. 29, based on anonymous sources, under the headline “A Defense Contractor Gave Their Personal AI Agent Access to a Secure Pentagon System.” It says an Accenture employee working as a contractor for the U.S. Department of Defense (DOD) connected a secure department system to a personal Claude Code account earlier this year. Sources told WIRED the employee gave the agent their own login credentials to automate writing code and setting up cloud servers. According to the sources, that gave the agent administrative access and let it bypass security controls enforced through the normal login process. The agent did not just suggest changes; it wrote code and committed it to the system on the employee’s behalf.

WIRED describes the system as a secure centralized system used by multiple intelligence agencies. Its coding environment runs on NIPRNet, the DOD’s internet-connected network for unclassified information. The system holds controlled unclassified information (CUI), a category covering information that is sensitive but not classified. According to the sources, the agent ran on the network for about two months before it was identified and the vulnerability could be patched.

Information reportedly sent out

According to WIRED, during that time the agent appears to have sent the employee’s login credentials, passwords and a map of NIPRNet to a third-party website. WIRED did not name the site, saying that doing so could increase the security risk. It wrote only that the National Institute of Standards and Technology (NIST) had previously identified the site as high risk. According to GBHackers and FourWeekMBA, NIST’s warning concerned a vulnerability that could let attackers inject instructions into a coding agent to steal credentials. There is no confirmation that this vulnerability was used in the incident.

It is not known how the information was sent, whether the third-party service retained the data or whether anyone obtained and misused it. WIRED also says there is no indication that attackers reached a classified or top secret system. The sources further allege that Accenture did not initially report the incident to the government at all. According to FourWeekMBA, DOD employees and federal contractors must report incidents within 72 hours of discovery.

The responses

In a statement, Accenture said: “We take all concerns seriously. These allegations were thoroughly and independently investigated and found to be without merit. We fully complied with our legal and contractual obligations.” The company did not say who conducted the investigation. The Defense Department said it does not comment on the status of its networks or systems.

Anthropic did not comment on the incident itself. It pointed WIRED to its usage policy, which tells users not to bypass security controls such as authentication systems, endpoint protection or monitoring tools without authorization. No official document on the incident has been published; the entire account rests on WIRED’s anonymous sources and has not been confirmed by the Defense Department.