By Hamza Faisal
OpenAI is inviting outside hackers to break its own systems. The company announced on November 6 that it has partnered with Hacktron, a cybersecurity firm that builds AI agents for code review, penetration testing and red teaming, to hunt for vulnerabilities across OpenAI's infrastructure — with the focus on the seams where its systems meet third-party services.
Under the partnership, OpenAI engineers and Hacktron's security researchers will jointly choose the highest-priority areas to test, rather than Hacktron working from a fixed list. OpenAI is also expanding Hacktron's access to its frontier models through Daybreak Red, its program for putting advanced AI in the hands of vetted defensive security researchers.
A partnership that started with a bug bounty
OpenAI says it first met Hacktron through its Security Bug Bounty program, where the firm's researchers found and disclosed vulnerabilities serious enough to catch the company's attention. Hacktron's research has exposed weaknesses in how applications, cloud infrastructure and identity systems interact — the same kind of boundary the new partnership is built to probe.
"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," Hacktron chief executive Zayne Zhang said in the announcement. He added that the firm is combining its offensive-security expertise with OpenAI's models to help secure OpenAI and the broader software ecosystem.
The seams are the target
The choice of focus is deliberate. OpenAI's systems increasingly connect to outside services — tools, data sources and other companies' software — and those connections are where many modern breaches begin. A flaw in a single integration can hand an attacker a path into the larger system, and testing those boundaries requires people who think like attackers, not just engineers checking their own code.
The Hacktron deal is one of several security moves OpenAI made in the same week. The company also brought in researchers from Cathedral for its next security sprint and raised its bug bounty rewards, framing the push as part of an effort to accelerate cyber defence with frontier AI.
Why it matters
OpenAI's announcement discloses no vulnerability, no finding and no number — it describes an arrangement for deeper, jointly scoped testing, not a result. The significance is structural: the company that builds some of the world's most capable AI models is handing outside researchers broader access to those same models and asking them to find the weak points before someone else does.
For an industry racing to connect AI agents to real-world tools and services, that is the right question to ask. Agents that can act on a user's behalf are only as safe as the integrations they touch, and the fastest way to find out where those integrations break is to let skilled attackers try. OpenAI has now hired them to do exactly that.



