AI-as-a-Service Botnet Routes Malicious Workloads Across Compromised Windows and Linux Hosts

The report analyzes an underground advertisement for the Mycelium Framework, an advertised AI-as-a-Service botnet that claims to combine credential harvesting, remote exploitation, persistence, kernel/rootkit components, and distributed AI tasking across compromised endpoints; although no source code or proof-of-execution was provided, the individual techniques have precedent and the integrated concept poses plausible, high-impact risks requiring expanded hunting for stolen API keys, anomalous GPU/CPU inference workloads, encrypted outbound tasking, and correlated browser-credential theft.