Researchers Find LLM Reasoning and AI Safety Warning Left Inside Unfinished 'TuxBot v3 Evolution' IoT Botnet

Security researchers at Palo Alto Networks' Unit 42 uncovered a previously unreported IoT botnet framework called TuxBot v3 Evolution that still carried telltale signs of AI-assisted development, including raw large language model reasoning comments and an AI safety disclaimer embedded directly in the code. The findings, detailed in reporting on the analysis, offer a concrete look at how attackers are beginning to lean on generative AI to build malware — and where those efforts still fall short.