Warden Stealer: a rapidly rising info-stealer targeting AI agent data

Gen Threat Labs has published a technical analysis of Warden Stealer, noting that this Rust-written MaaS info-stealer is among the first to steal local data from AI assistants and agent coding tools, with targets including Claude, Codex, Grok, and Cursor. It can steal access and refresh tokens, credentials in MCP configurations, prompt history, and conversation databases, thereby gaining both account access and context on the value of the projects behind that data. Earlier collection of AI agent data came from operator-defined custom rules; starting with version 1.9 (announced September 29, 2026), official support was added for collecting AI coding agent tokens. It has been promoted on underground forums since August 2026, and the author says roughly 110 customers are currently using it; Gen says it has become one of the most popular info-stealers in its user base.