← All articles
OpenAI hires security firm Hacktron to test its own systems
AI News

OpenAI hires security firm Hacktron to test its own systems

OpenAI has formally brought in Hacktron, the firm whose researchers once breached its systems, to hunt for weaknesses in third-party connections.

OpenAI announced on 6 November 2026 a partnership with Hacktron, the security firm whose researchers once breached OpenAI's own systems. This time they are being brought in officially to find weaknesses. The main focus is the connection points between OpenAI systems and third-party services.

1. What was announced

OpenAI engineers and Hacktron's security researchers will choose the highest-priority testing areas together. The first target is where OpenAI's platforms connect to outside services, such as forums, single sign-on providers and SaaS connectors.

As part of the deal, Hacktron gets expanded access to OpenAI's frontier models through Daybreak Red, OpenAI's restricted program for offensive security work. No contract value or duration has been disclosed.

2. From bug hunter to partner

The relationship started in OpenAI's Security Bug Bounty program. In September, NBC News reported that Hacktron researchers chained two previously unknown flaws to reach OpenAI employees' ChatGPT accounts.

One flaw was in Discourse, the third-party forum software OpenAI uses, and the other was in how OpenAI verified its staff. According to NBC, the work took under 72 hours in late July and caused no damage. OpenAI confirmed the report, said the issues had been patched, and paid a $6,500 bounty, a figure given in Hacktron's own write-up and repeated by NBC and NewsBytes.

3. Why third-party connections are the target

OpenAI frames the risk directly: as AI agents connect services in new ways, the trust boundaries between those services matter more and more, and they are hard to test from outside without context.

Hacktron builds AI agents for code review, penetration testing and red teaming. Its research has uncovered weaknesses in how applications, cloud infrastructure and identity systems interact. That is exactly the kind of path used to reach employee accounts last month.

4. It follows a week of security moves

The partnership was announced a day after OpenAI opened its new GPT-6 Cyber model to selected Daybreak Red participants, according to reporting by HuggingNews. The model itself was released on 5 November.

Earlier the same week, OpenAI also said Cathedral's researchers will join its next security sprint, and raised its bug bounty rewards, with the top payout reaching $172,000 according to Help Net Security on 3 November.

OpenAI's official announcement card for the GPT-6 Cyber model
OpenAI's official card for the GPT-6 Cyber security model, opened to Daybreak Red participants. Image: OpenAI.

5. What both sides say

Hacktron's chief executive, Zayne Zhang, said the future of cybersecurity will be defined by whoever best combines human creativity with AI-driven capabilities, not by who has the largest security team.

OpenAI said working alongside Hacktron will help it test those connections more deeply, and that what it learns will be used to strengthen its defenses.

What it means for you

This attack pattern is not specific to AI companies. Most businesses, including small businesses in Malaysia, rely on forums, ticketing systems, single sign-on providers and SaaS connectors they did not write. Weaknesses usually sit at the seams between those services.

  1. Audit your outside connections. List the third-party software that can connect to your core systems, and know who authenticates access on each side.
  2. Do not rely on one security layer. The July path combined an outside software flaw with an internal verification weakness. Separate layers are harder to chain.
  3. Test before someone tests you. Regular penetration testing at integration points costs less than an incident.
  4. Consider a bug bounty if you run a product. Hacktron's story shows a good bounty relationship can turn into paid work with privileged access.

Sources