Hacking OpenAI

Evidence record

Primary narrative and timeline for the Hacktron/OpenAI/Discourse chain. Researcher-authored disclosure; broader connected-service reach is partly theoretical and should be separated from demonstrated actions.

Read the original source ↗

Claim-level citations (8)

• supportsHacktron: image processing, SSO and connected development authority: Hacktron reports a July 25 chain from Discourse image processing through an OpenAI SSO flaw into employee ChatGPT/Codex accounts, demonstrated with a harmless internal pull request.

  Intro; disclosure timeline; Opus 5 Released

• supportsHacktron: image processing, SSO and connected development authority: The authors identify Opus 4.8 and Opus 5. They say they did not read internal code; access to other connectors was potential rather than demonstrated.

  Background; Opus 5 Released

• supportsHacktron: image processing, SSO and connected development authority: The sub-$3,000 token figure covers broader multi-company research. The $6,500 award covered the OpenAI-side finding; the quoted scope clarification excludes testing the Discourse-hosted forum from that bounty.

  Costs of finding these vulnerabilities; disclosure timeline item 9

• supportsHacktron: image processing, SSO and connected development authority: The researchers demonstrated internal development authority by having Codex open a harmless pull request, while stating that they did not read internal OpenAI source code.

  OpenAI employee account/Codex proof-of-access section

• supportsHacktron: image processing, SSO and connected development authority: OpenAI fixed the account-takeover path on July 25, and Discourse subsequently published a HEIF advisory and image-processing mitigations.

  timeline; OpenAI fix confirmation

• supportsHacktron: image processing, SSO and connected development authority: Hacktron says Claude materially accelerated exploit development, including an autonomous loop against a researcher-controlled Discourse instance, but skilled human guidance remained important and the work was not completely autonomous.

  sections describing Opus 4.8/Opus 5 exploit development and the /goal loop

• supportsHacktron: image processing, SSO and connected development authority: The disclosure discusses broader potential access through connected services, but it does not establish that those additional services were actually accessed.

  discussion of connected applications and stated stopping point

• supportsHacktron: Security research organization behind D01; commercial AI-security platform. Role is scoped to the cited page/date.

  Intro; Team / About