This payment addresses a vulnerability that could have allowed unauthorized code execution within user
This payment addresses a vulnerability that could have allowed unauthorized code execution within user sessions.
Key points
- The bug allowed attackers to run code in the logged-in sessions of victims.
- Exploitation required victims to open a malicious link shared by the attacker.
This payment addresses a vulnerability that could have allowed unauthorized code execution within user sessions. The disclosure involves a specific bug found in the company's consumer-facing products.A security researcher found a way to run code in other people's ChatGPT accounts by sending them a bad link.
What changed
The flaw specifically affected how ChatGPT handles shared conversation links. Before this report, the vulnerability remained undisclosed to the public. The payment confirms that OpenAI processed a bounty reward for the finding.
How it works
The vulnerability allowed an attacker to run code in the logged-in session of any user who opened a malicious ChatGPT share link. According to the summary provided by Techmeme, the exploit required victims to open the link.reported Once opened, the attacker's code would execute within the context of the victim's active session.
Evidence: reported
Researchers are paid $5K for reporting a bug to OpenAI that would let attackers run code in logged-in sessions of all who opened the hacker's ChatGPT share link (Swati Khandelwal/The Hacker News)aifleet-techmeme
What the evidence means
The scope of the attack was limited to users who actively opened the specific malicious link. It did not affect users who did not interact with the compromised URL. The report confirms that the flaw existed in the sharing mechanism of ChatGPT conversations. No details were provided about the duration the bug remained active before discovery.
Cost and market position
This bounty represents the value the company places on preventing unauthorized code execution in its ChatGPT product. For product teams, this demonstrates that consumer AI platforms face significant session hijacking risks.
Why it matters
- Product teams must evaluate how shared links in their applications handle session integrity.
- This incident shows that malicious links can be used to exploit web flaws.
- OpenAI’s response indicates active monitoring of its consumer product security posture.
Topics: OpenAI, ChatGPT, security, vulnerability, bug bounty
Updated
Then / now / next
ChatGPT share links were assumed to be safe for sharing conversations.
A vulnerability allowing code execution in victim sessions has been disclosed.
Teams should monitor for further details on the exploit and any patches.
Story timeline
- Then ChatGPT share links were assumed to be safe for sharing conversations.
- 31 Oct 2026 Researchers are paid $5K for reporting a bug to OpenAI that would let attackers run code in logged-in sessions of all who opened the hacker's ChatGPT share link (Swati Khandelwal/The Hacker News)
- Now A vulnerability allowing code execution in victim sessions has been disclosed.
- Next Teams should monitor for further details on the exploit and any patches.
The strongest doubt
Counterpoint from the sources
The source material does not provide any opposing views or doubts about the severity of the bug.
For readers
Developers
- Review how your application handles shared links and session tokens.
- Ensure that shared URLs do not execute arbitrary code on the client side.
Business
- Consider the cost of security bounties as part of product risk management.
- Evaluate the reputational risk of session hijacking vulnerabilities.
Researchers
- Submit findings through official channels to receive bounties.
- Focus testing on social engineering vectors like malicious links.
Ask the article
Local keyword search of this article; no generated answers.Did this article give you the evidence you needed?
Your choice stays on this device. No public tally.Questions readers ask
Who was affected by this bug?
Users who opened a malicious ChatGPT share link were at risk of having code run in their session.
Is the bug fixed?
The source material does not confirm if a patch has been released or if the bug is fully resolved.
Connected stories
Reading receipt
- The bug allowed attackers to run code in the logged-in sessions of victims.
- Exploitation required victims to open a malicious link shared by the attacker.
Still open: Has OpenAI released a technical advisory or patch for this vulnerability?
Still open: How long was the vulnerability present before the researcher reported it?
