Executive Summary
Attackers are currently targeting macOS users through a sophisticated 'ClickFix' campaign that leverages trusted services like Google Ads and Google Sites. By impersonating developer tools such as OpenAI Codex and Claude, threat actors persuade victims to manually copy and paste malicious commands into their Terminal under the guise of an application installer. This technique bypasses traditional download warnings and places the burden of execution on the user.
The infection chain is a multi-stage process involving shell-script loaders, obfuscated Zsh stages, and final Mach-O payloads. Analysis shows significant overlap with the Atomic macOS Stealer (AMOS) delivery framework, including the use of specific telemetry endpoints and universal Mach-O structures designed to run on both Intel and Apple Silicon architectures.
This campaign highlights an evolving trend where attackers abuse legitimate AI-branding and no-code hosting platforms to establish trust. The impact for organizations includes the potential for broad credential theft and sensitive data exfiltration from macOS endpoints, particularly targeting developers and technical staff who frequently use Terminal and AI tools.