Accenture Contractor Allegedly Connected Personal Claude Code to US DoD Network
Accenture Contractor Linked Personal Claude AI Agent to Pentagon System
A WIRED investigation revealed that an Accenture contractor allegedly connected a personal Anthropic Claude Code agent to the US Department of Defense's NIPRNet production environment for about two months, and allegedly leaked credentials and network topologies to a third-party website. The agent reportedly held administrator privileges and could deploy code and cloud servers. Accenture responded that an independent investigation found the allegations to be entirely unsubstantiated. There is currently no public evidence indicating that classified networks were compromised or breached.
The incident highlights the data leakage risks posed by uncontrolled permissions of personal coding agents in sensitive network environments, serving as a security warning for organizations deploying and governing AI tools.
Source:Hacker News · AI · ibtimes.co.uk