VelLMes, a high-interaction AI based deception framework

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F24%3A00380909" target="_blank" >RIV/68407700:21230/24:00380909 - isvavai.cz</a>

Result on the web

<a href="https://www.blackhat.com/eu-24/arsenal/schedule/index.html#vellmes-a-high-interaction-ai-based-deception-framework-42075" target="_blank" >https://www.blackhat.com/eu-24/arsenal/schedule/index.html#vellmes-a-high-interaction-ai-based-deception-framework-42075</a>

DOI - Digital Object Identifier

—

Result language

angličtina

Original language name

VelLMes, a high-interaction AI based deception framework

Original language description

is based on new deception research that uses fine-tuned and trained LLMs to create high-interaction honeypots that look exactly like your production servers. When attackers connect to a VelLMes SSH, they can not distinguish it from a real Linux shell. The LLM creates in real time, and depending on the commands of the attacker, the complete structure of the simulated computer, including file contents, output of all commands, connection to the Internet (simulated), users, and more.

Czech name

—

Czech description

—

Type

O - Miscellaneous

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Publication year

2024

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů