Attackers can craft a malicious prompt that coerces the language model into executing arbitrary JavaScript in the context of the web page.
{
"cna_assigner": "KoreLogic",
"cwe_ids": [
"CWE-79"
],
"osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2024/6xxx/CVE-2024-6706.json"
}