Disclosed weakness · no CVE recorded

ChatGPT shared-link MathLive message injection

OpenAI · ChatGPT web

First documented disclosure · Review checkpoint

Summary

A shared conversation could access private chats.

Severity and exploitation evidence

Unknown · numeric CVSS score not recorded. Severity source.

Exploitation not established: Controlled research; malicious use unestablished.

Exploitation evidence source

Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.

How the weakness works

Unchecked MathLive messages inserted trusted-page HTML; an unsandboxed frame bypassed script restrictions.

Conditions required

Signed-in user opens a crafted share; Preview activation is required but could occur automatically.

Potential impact

Authenticated session access and private-chat disclosure.

Affected versions and configurations

Pre-fix web frontends; builds unspecified.

Fix and mitigation status

Both frontends fixed October 26, per Parameter.

Practical next steps

Provider fix deployed; validate cross-frame senders and sandbox generated documents.

Operating lesson

A preview needs secure messaging boundaries.

Evidence and limits

Researcher reports OpenAI P1 priority, not CVSS. No CVE named.

Related incident and research records

Related records can include laboratory demonstrations; a relationship does not itself establish exploitation.

Sources