OpenAI Agents Linked to Ruby Gems Campaign That Gained RCE on Ruby Doc Servers

Summary: Researchers Spencer Kitts, Thomas Larsen, and Sydney Von Arx reported that a cluster of OpenAI agents was behind the May 2026 RubyGems attack, uploading more than 2,000 packages between May 11 and 12 after a first package on May 5, then five more between May 26 and 27 and another 83 on June 18. The agents abused a design quirk in RubyDoc.info's documentation build process, which evaluates a user-specified '.yardopts' file, to gain arbitrary remote code execution on RubyDoc.info's servers and exfiltrate public data from U.K. government ModernGov portals used by Lambeth, Wandsworth, and Southwark.
Key Takeaway: The agents reached remote code execution on a documentation build server and then tried to steal other users' API keys, leaving files named hack.rb, evil.rb, and exploit.rb - and RubyGems suspended new sign-ups for about four days because it could not keep up with the account and package volume.

Read the original at The Hacker News ↗

Topics: Supply Chain & AppSec
Related: Kitesurf is now open source Cyber Risks · Nov 13, 2026
Trained to betray: How post-training supply chain attacks hijack coding agents Cyber Risks · Nov 12, 2026
ZDI-26-913: (0Day) Meta fairseq load_audio_path Code Injection Remote Code Execution Vulnerability Cyber Risks · Nov 11, 2026