The US Army has launched Project Gatekeeper to help commercial developers identify and fix security flaws in their software before seeking approval to use it on army networks.
Introduced by the Army’s Office of the Chief Information Officer, the system provides isolated environments where companies can submit code and applications for automated security checks.
Developers receive feedback on vulnerabilities and instructions on how to address them.
The platform is intended to simplify a process that, according to the army, has often required companies to hold defense contracts, obtain military sponsorship, or secure access credentials to have their software evaluated.
Approval timelines could stretch to six to eight months or longer.
Gatekeeper will initially be available to vendors with a Department of Defense Common Access Card.
The army said access through Login.gov would follow, allowing commercial users to authenticate without those military credentials, but did not provide a launch date.
Automated Checks Support Approval Decisions
Gatekeeper automatically scans software for security vulnerabilities and gives developers guidance on how to fix them.
It also generates standardized reports documenting the test results and compliance with US Army security requirements.
Officials can use that evidence when deciding whether to grant an authority to operate, the formal approval needed to run a system on US Army networks.
Acting Army Chief Information Officer Gabe Chiulli said the platform could cut initial evaluation cycles from months to minutes.
Officials must still review the results, assess any remaining vulnerabilities, and decide whether the software can be authorized for use.
To carry out those checks, Gatekeeper runs applications in isolated, container-based environments.
These keep the software separate from operational systems and data while allowing developers to test it against army security standards.
The US Army also plans to add advanced AI security models to support its existing scanning tools.
These models would help bring findings from different checks into a consolidated report and identify which vulnerabilities developers should address first.