[ad_1] Anthropic’s Frontier Red Team spent the past year teaching AI agents to behave like professional DeFi attackers. The agents learned to fork blockchains, write exploit scripts, drain liquidity pools, and pocket... Read more »