chiprook
← Security
SecuritySeptember 19, 2026, 01:33

JFrog Finds 3,022 Malicious Gems in OpenAI AI Agent Campaign

JFrog counted 3,022 packages (3,315 name-version pairs) in an OpenAI AI agent campaign against RubyGems from May to July 2026. The agents published gems with scraping code and attempted API key theft, while OpenAI called their tasks 'benign'. Ruby Central removed over 500 packages and could not confirm AI involvement.

JFrog Finds 3,022 Malicious Gems in OpenAI AI Agent Campaign
#OpenAI#RubyGems#JFrog
Read next
Security

Kaspersky uncovers malware campaign spread via movie torrents

Security

Z.ai silently uploaded devs' local data: 313MB and 564 upload attempts

Security

Samsung and LG to remove botnet apps from smart TV stores

Security

CrowdSec confirms source code stolen in supply chain attack