Incident report · Data exfiltration
A Shared ChatGPT Link Could Run Attacker Code in Your Session and Read Private Chats
Security firm Parameter discovered a critical vulnerability in ChatGPT's shared link functionality. A malicious shared link could execute attacker code in a user's logged-in session, allowing access to their private conversations. The flaw was reported to OpenAI on October 18, 2026, and fixed eight days later, receiving a P1 rating. The extent of data exfiltration is unknown.
- Discovered
- Oct 18 2026
- Failure mode
- Data exfiltration
- Source
- The Hacker News · Oct 31 2026
- Impact
- Users' private conversations could be read, attacker code execution in user session.
Summary from the public record. The failure mode’s analysis follows.
Every answer was harmless on its own. Together they were the whole table.
How Fencio prevents this
Sensitive data and an outbound channel ended up in the same context. The agent did not need to be malicious. It only needed to be convinced that sending the data somewhere was part of the job.
Fencio tracks sensitive data as it moves through an agent session and checks every outbound path, from links and images to emails and API calls. When classified data is about to leave, or a series of answers adds up to something the requester is not entitled to see, the response is blocked or redacted.
Runtime policy example · Data exfiltrationLive
- 13:44:56respond(aggregate=comp.bands)Cumulative disclosure over thresholdBlock
- 13:47:13respond(policy.summary)Public classificationAllow
- 13:50:30render(img src=external?d=...)Data encoded in outbound urlBlock
Say yes to the AI your business wants.
Book a 30-minute call. Bring your agents and your rules.