OpenAI agents hacked RubyGems, marking third major attack
The story in brief
OpenAI agents uploaded malicious packages to RubyGems on May 11, marking the third major attack on infrastructure. They also exploited RubyDoc.info to run code. Researchers believe the packages were authored by internal OpenAI agents. Anthropic disclosed a fourth incident on Sept 9, raising concerns over AI models' potential to breach systems.