CRACKEN JOINS OPEN SECURE AI ALLIANCE TO ADVANCE PROACTIVE CYBERSECURITY

CRACKEN JOINS OPEN SECURE AI ALLIANCE TO ADVANCE PROACTIVE CYBERSECURITY

Cracken has joined the Open Secure AI Alliance (OSAIA), a Linux Foundation project. The Alliance develops open tools, shared standards and defensive practices that help organizations secure software and AI agents.

Cracken’s platform was built and field-tested in active cyber warfare. The company will bring its expertise in sovereign and safe adversarial-grade AI and proactive cybersecurity to the Alliance, with plans to propose research and open-source projects that help defenders secure any attack surface.

As a member, Cracken will work alongside some of the world's leading AI and cybersecurity companies to support the Alliance's mission of building an open defensive stack that gives defenders capabilities they can inspect, adapt, trust and control.

“We’re happy to have been invited to join the Alliance. Its vision aligns closely with how we see the future of proactive cybersecurity. Attackers don’t care about AI guardrails, governance or approval processes, so defenders need to stay one step ahead by finding and validating exploitable weaknesses across attack surfaces before they become breaches. We look forward to contributing our experience and research to this collective effort.”

— Artem Sorokin, CEO and Founder, Cracken.

Cracken’s existing open-source projects include BlackSea, a honeypot that deploys lures resembling real assets to detect, profile and disrupt autonomous AI attackers before they reach real systems. The second is RedLineBench, an open benchmark that measures the tension between offensive-security capability and AI safety alignment. Across 153 offensive-security tasks, it separately scores whether a model refuses a request and, when it answers, whether the answer is technically valid and usable. Both reflect Cracken’s approach of publishing adversarial research and building tools that defenders can use and adapt.