WardenStealer
Latest
Infostealers Are Stealing ChatGPT and Claude Tokens That Survive Password Resets
SpyCloud reports recapturing more than 130,000 unique WardenStealer infections since August 2026, with infostealers now targeting AI developer environments. The malware harvests authentication files for ChatGPT and Claude, which can preserve unauthorised access after a password reset. The research describes an expansion of commodity credential theft rather than a flaw in the AI models themselves.
WardenStealer and Aoki infostealers target ChatGPT, Claude and Codex tokens that survive password resets
The Windows infostealers WardenStealer and Aoki are harvesting OAuth tokens, cookies and API keys from AI tools including ChatGPT, Claude Code, Codex, GitHub Copilot and Cursor, according to SpyCloud. Access and refresh tokens may remain valid after a password reset, allowing attackers to act as authenticated users without a fresh MFA challenge. SpyCloud reported more than 130,000 unique WardenStealer infections since the family was first seen on 3 August 2026, and stolen AI access is being sold cheaply for LLMjacking.