CloudSEK Exposes AI-Agent-Driven Crypto Theft Campaign Behind Mass Wallet and Credential Compromise

CloudSEK, a cybersecurity firm, disclosed on August 19 a new offensive operation driven by AI agents, saying an exposed adversary open directory revealed an autonomous crypto-theft campaign. The findings point to mass wallet and credential compromise.\n\nThe report is titled AI-Agent-Driven Offensive Operation, framing the discovery as an autonomous campaign rather than a conventional manually run heist.\n\nThe accidental exposure of the adversary's open directory gave researchers a window into the automated attack chain behind the thefts.\n\nUnlike phishing or theft campaigns that rely on human involvement, this operation appears to have been executed autonomously by AI agents, raising both the scale and the speed of the attacks.\n\nFor crypto users, the mass compromise of wallets and credentials signals that asset risk is shifting from individual mistakes to an expanding automated attack surface.\n\nFor the industry, the disclosure is a concrete example of AI capabilities being turned to offensive use, validating long-standing warnings about the weaponization of AI agents.\n\nWhat to watch next is whether CloudSEK publishes technical indicators and the scope of affected users, and whether AI-driven attacks of this kind spread beyond the crypto sector.