OpenAI expands security testing partnership with Hacktron

OpenAI has announced a partnership with Hacktron to identify vulnerabilities in OpenAI systems, with joint work focused on high-priority areas and on how those systems interact with third-party services. As part of the arrangement, OpenAI is expanding Hacktron’s access to its frontier models through Daybreak Red.

OpenAI said its engineers and Hacktron’s security researchers will select the areas they test together. The initiative is intended to help the company investigate connections between services more deeply and apply the findings to strengthen its defenses.

Testing the boundaries between interconnected systems

Hacktron first came to OpenAI’s attention through the Security Bug Bounty program. OpenAI cited the impact of vulnerabilities that Hacktron researchers had found and responsibly disclosed through that channel.

The Hacktron team builds AI agents for code review, penetration testing and red teaming. Its research has examined vulnerabilities arising from interactions among applications, cloud infrastructure and identity systems. OpenAI said this work illustrates how security engineers can use AI to move faster and pursue more ambitious research.

That emphasis reflects a practical security problem: as AI agents connect services in new ways, the boundaries between those services warrant deeper testing. OpenAI is pairing Hacktron’s offensive-security expertise with access to frontier models in order to examine those interactions.

Part of a wider OpenAI security effort

The partnership follows other measures announced by OpenAI. The company said Cathedral researchers will work alongside its teams during the next security sprint, and it raised bug bounty rewards to encourage independent researchers to seek weaknesses that internal testing may not detect.

OpenAI has also described an initiative in which OpenAI support for securing open-source software supports the protection of open-source software, underscoring its stated focus on using advanced AI capabilities in cyber defense.

Implications for security programs

Hacktron CEO Zayne Zhang said the future of cybersecurity will depend on combining human creativity with AI-driven capabilities. He said access to frontier cyber models can help researchers uncover complex vulnerabilities at the boundaries of interconnected systems.

OpenAI thanked Hacktron and other independent researchers for reporting weaknesses through its Security Bug Bounty program. For businesses, the immediate implication is to give priority to testing service integrations, cloud infrastructure and identity flows, while maintaining clear vulnerability-reporting channels and remediation processes.