OpenAI partners with Hacktron to find security flaws in connected services

OpenAI is bringing an outside security team closer to its systems to examine how they connect with other services. In a November 6, 2026 announcement, the company said Hacktron researchers would jointly choose testing priorities with its engineers. The promise is deeper scrutiny of those connections; the announcement describes planned work, not new vulnerability findings.

Testing the connections between systems

The testing will focus on interactions between OpenAI’s systems and third-party services. OpenAI frames those connections as an increasingly important security target as AI agents link services in new ways. The stated goal is to examine the connections more deeply and use the findings to strengthen its defenses.

Hacktron’s previous research fits that focus. According to OpenAI, its researchers have uncovered vulnerabilities involving interactions among applications, cloud infrastructure and identity systems. That work concerns the boundaries between connected systems, rather than only weaknesses within a single application.

The relationship began through OpenAI’s Security Bug Bounty program. OpenAI says Hacktron researchers discovered vulnerabilities and disclosed them to the company, which was impressed by their impact. The new arrangement adds joint selection of high-priority testing areas to that earlier relationship built around reporting flaws.

The future of cybersecurity won’t be defined by who has the largest security team, but by who can best combine human creativity with AI-driven capabilities.

Zayne Zhang, CEO of Hacktron, in OpenAI’s partnership announcement

More model access for the researchers

As part of the partnership, OpenAI is expanding Hacktron’s access to its most advanced models through Daybreak Red. The access change is tied directly to the joint security work: OpenAI is supplying models to the researchers who will help examine weaknesses in its own systems.

Hacktron already builds AI agents for code review, penetration testing and red teaming—work that examines software and challenges systems to find weaknesses. OpenAI describes the team’s approach as combining security engineers’ skills with AI, allowing them to move faster and pursue more ambitious research. That is OpenAI’s assessment, not a measured result from the new partnership.