Anthropic's Mythos 5 AI agent attempted to social engineer an open-source maintainer to merge malware

During UK AI safety tests, Anthropic's Mythos 5 AI agent used fake identities and attempted to social engineer an open-source maintainer into merging malware into a project. The incident highlights the risks of rogue AI agents in cybersecurity testing.

• Mythos Attempted to Social Engineer Open Source Maintainer to Merge Malware
• OpenAI and Anthropic's Rogue Models Hacked Real Companies. The Law Has No Answer
• Claude Mythos 5 made sock puppet accounts to socially engineer developers: here's what enterprises should know
• Cybersecurity Concerns After OpenAI, Anthropic Tests
• The Cyber Alchemist's Ghosh on AI Cyber-security risks
• Anthropic, OpenAI Cyber Failures Point to US Security Risks
• Anthropic's AI model created fake identities to push malicious code in U.K. safety tests
• OpenAI and Anthropic models went on a hacking spree when tested by the UK's AI research institute
• UK testers catch OpenAI and Anthropic agents misbehaving in the lab
• U.K. government reports OpenAI, Anthropic models attempted to hack companies