# OpenAI expands security testing with Hacktron for third-party service links

Digest AI · Policy & Regulation · published 2026-11-06T18:41:03Z

Canonical: https://digestai.news/story/openai-expands-security-testing-with-hacktron-for-third-party-service

## Summary

OpenAI and Hacktron announced a partnership on November 6, 2026, to test vulnerabilities in OpenAI’s systems, focusing on connections between third-party services. The collaboration builds on Hacktron’s existing work through OpenAI’s Security Bug Bounty program, where the firm previously uncovered critical flaws in OpenAI’s systems, including a 2026 incident where researchers chained two vulnerabilities to compromise multiple ChatGPT accounts and access internal repositories. The new partnership will involve joint testing of how AI agents interact with linked services, with lessons applied to strengthen OpenAI’s defenses against weaknesses in application, identity, and infrastructure integrations.

Hacktron’s access to OpenAI’s Daybreak Red program—designed for authorized vulnerability research—will allow deeper testing of frontier models in controlled environments. OpenAI’s Daybreak framework includes safeguards like identity verification, monitoring, and restricted-use policies to mitigate risks. The partnership emphasizes a balance between expanded testing capabilities and human oversight, ensuring that while AI agents may automate parts of the process, skilled researchers retain critical decision-making authority.

## Key points

- OpenAI and Hacktron partner to test vulnerabilities in third-party service integrations, expanding Hacktron’s Daybreak Red access
- Hacktron previously chained two flaws in July 2026 to breach ChatGPT accounts, leading to a $6,500 bounty from OpenAI
- Daybreak Red provides controlled access to frontier models for authorized security research with strict oversight

## Why it matters

This partnership signals OpenAI’s growing focus on securing its expanding ecosystem of third-party integrations, a critical area as AI agents increasingly connect services. For the industry, it highlights the need for rigorous testing of interconnected systems to prevent breaches that could expose sensitive data or internal workflows.

## Sources

1. [OpenAI Partners With Hacktron to Test Third-Party Service Connections](https://unite.ai/openai-partners-with-hacktron-to-test-third-party-service-connections) (Unite.AI, 2026-11-06)

Part of the developing story: [OpenAI's Quest for AI Security](https://digestai.news/thread/openai-raises-bug-bounty-rewards-to-172-000-and-plans-security-sprint) (3 stories)

## Cite

Digest AI, "OpenAI expands security testing with Hacktron for third-party service links", 6 November 2026, https://digestai.news/story/openai-expands-security-testing-with-hacktron-for-third-party-service

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/openai-expands-security-testing-with-hacktron-for-third-party-service.json