Security researchers have leveraged Anthropic's cutting-edge AI model, Mythos, to uncover a new security bypass technique within Apple's macOS 26.4.1. This breakthrough enabled them to escalate privileges and sidestep Apple's key memory safety safeguard, MIE. Initiating from a standard local account, the research team ingeniously combined two vulnerabilities to construct a complete attack chain within a mere five days, ultimately acquiring root shell privileges. Presently, Apple is in the process of evaluating the reported vulnerabilities, and the research team has opted to withhold the full technical specifics for the time being.
