Even after a fix was issued, lingering prompt injection risks in GitLab’s AI assistant might allow attackers to indirectly deliver developers malware, dirty links, and more.

from darkreading https://ift.tt/DcaNBTw
via IFTTT