A severe authorization vulnerability was recently discovered in Schemata, an AI-powered virtual training platform serving the United States Department of Defense.
Security researcher Alex Schapiro, utilizing the open-source AI hacking agent Strix, identified a critical lack of API authorization. Backed by Andreessen Horowitz, Schemata holds active government contracts to provide immersive 3D simulations for various military branches.
This zero-auth flaw allowed unprivileged accounts to access sensitive military training materials and personnel records across different tenant boundaries, highlighting a significant operational security risk.
Zero-Auth Vulnerability
The automated Strix agent established a low-privilege baseline to map the Schemata application’s reachable API surface. By observing normal application behavior and client-side route references, the tool replayed high-value collection endpoints using a standard, unprivileged session.
The system failed to isolate tenants, returning data for the entire platform rather than restricting it to the specific user or organization. The API lacked any meaningful organizational scoping or permission checks.
Furthermore, the absence of authorization checks on write-enabled routes meant attackers could potentially manipulate or permanently delete critical training infrastructure.
The exposed endpoints provided direct access to an alarming volume of sensitive intelligence. Unprivileged accounts could query the complete user directory to extract full names, email addresses, and enrollment data.
Crucially, this included the specific military base deployments of active U.S. service members, leaving them highly vulnerable to targeted phishing or doxing campaigns. The vulnerability also exposed metadata and direct AWS S3 links for hundreds of confidential modules for military operations.
These compromised assets included proprietary 3D naval maintenance training materials. They restricted Army field manuals concerning the tactical deployment of explosive ordnance.
The vulnerability disclosure timeline illustrates significant challenges in security reporting. Strix researchers first contacted Schemata on December 2, 2025.
The company leadership’s initial response incorrectly assumed the disclosure was an attempt to solicit a bug bounty payment. Despite researchers clarifying their intent and sending multiple follow-up warnings about the critical exposure, the vulnerability remained live for nearly 150 days.
Schemata finally acknowledged the exposed endpoints and successfully patched the system on May 1, 2026, immediately before the scheduled public disclosure.
Organizations maintaining active Department of Defense contracts operate under strict federal regulations, including DFARS 252.204-7012 and CMMC requirements for handling Controlled Unclassified Information.
A platform that serves military data without a functional API authorisation layer constitutes a foundational cybersecurity failure. Government partners using Schemata should request access logs as soon as possible to verify data integrity and assess potential exposure during the vulnerability window.
Software developers must prioritize robust tenant isolation, continuous automated security validation, and maintain open channels for independent security researchers.
Follow us on Google News, LinkedIn, and X to Get Instant Updates and Set GBH as a Preferred Source in Google.
