OpenAI Makes Hacktron a Formal Security Testing Partner: Joint Focus on Third-Party Service Connections, With Expanded Daybreak Red Model Access
OpenAI announced on November 6 that it has partnered with offensive security firm Hacktron to find vulnerabilities in OpenAI's own systems. Under the announcement, OpenAI engineers and Hacktron security researchers will jointly select high-priority areas to test, with a focus on how OpenAI's systems interact with third-party services. As part of the partnership, OpenAI is expanding Hacktron's access to its frontier models through Daybreak Red, its restricted tier that provides purpose-trained cybersecurity models for authorized vulnerability research, exploit validation and security testing.
OpenAI said it became familiar with Hacktron through its Security Bug Bounty program and was impressed by the impact of the vulnerabilities the firm's researchers discovered and disclosed. Hacktron builds AI agents for code review, penetration testing and red teaming, and its research has uncovered vulnerabilities in interactions among applications, cloud infrastructure and identity systems. Hacktron chief executive Zayne Zhang said in the announcement that the company is excited to deepen its partnership with OpenAI, combining its offensive security expertise with OpenAI's frontier models to help secure OpenAI and the broader software ecosystem.
An earlier Hacktron disclosure forms the backdrop. In its September 13 account, Hacktron said its researchers chained two vulnerabilities on July 25 — one in the image-processing path of the Discourse community forum OpenAI uses, and one in OpenAI's single sign-on — to compromise multiple OpenAI employees' ChatGPT accounts, creating potential access to internal repositories and other connected services. Hacktron said OpenAI confirmed a fix roughly 14 hours after the report and paid a $6,500 bounty on September 1. OpenAI's partnership announcement itself discloses no new vulnerability or test result; it sets out the testing scope and the arrangement.
Source: OpenAI
Published by GSM · USA (original Chinese edition)
Photo: 图：OpenAI 标志。Wikimedia Commons，公共领域（PD-textlogo）。 / Photo: OpenAI logo. Wikimedia Commons, public domain (PD-textlogo).
Nature Communications Paper: A Single Dose of Propofol Suppresses Alcohol Drinking in Mice for About Six Days via Lateral Habenula Activation2026-11-08
Google Cloud Z4M Storage-Optimized Machines Reach General Availability: Up to 168,000 GiB of Local SSD, 400 Gbps Cloud RDMA and 15.6 Million Random-Read IOPS2026-11-08