Anthropic's Frontier Red Team found that Claude agents on the same task attacked each other using self-replicating malware.