A New Hacking Tool Can Put the Attacker's AI Bill on the Victim's Own Account

Cisco Talos reported Oct. 29, 2026, that it has found a Windows malware framework in active use whose built-in connection to OpenAI's paid programming interface falls back to the victim's own account. Talos says that when the attacker supplies no key of their own, the implant reads the OpenAI key configured on the infected machine and uses it, so the victim's payment method funds the attacker's AI usage and OpenAI's access logs record the victim's key as the one that sent the requests.

The framework, which Talos calls NIK C2, lets an operator dump stored passwords, stream the victim's keystrokes and then ask a model to analyze the harvested data inside the same compromised session, rather than pulling it back to the attacker's own machine first.

Talos says the implant, written in the Go programming language, also takes saved Wi-Fi passwords, browser login databases and crypto-wallet files, can capture the screen, camera and microphone, and can be ordered to lock the desktop and show a ransom demand payable in Bitcoin or USDT, a dollar-pegged crypto token. The company counted more than 52 commands the framework's control panel can issue. Talos says the screen-lock module encrypts nothing: it covers the display and restores itself after a restart.

Talos says it observed a live server hosting the framework during its research, carrying a security certificate whose name also appeared inside the implant file. Watching that server live, the company says, is what let it assess with moderate confidence that NIK C2 is a working framework and not a proof of concept. The company reports at least three distinct versions, all built with Go 1.23.4, a release from December 2024. That places the samples it examined no earlier than late 2024.

The analysis is Cisco's own threat-research work, published under the name of Talos researcher Chetan Raghuprasad, and carries no outside confirmation. Talos names no country, no group and no victim.