Attackers can craft a malicious prompt that coerces the language model into executing arbitrary JavaScript in the context of the web page.
The product does not neutralize or incorrectly neutralizes user-controllable input before it is placed in output that is used as a web page that is served to other users.
Link | Tags |
---|---|
https://korelogic.com/Resources/Advisories/KL-001-2024-005.txt | third party advisory exploit |
http://seclists.org/fulldisclosure/2024/Aug/3 | |
http://www.openwall.com/lists/oss-security/2024/08/08/6 |