LLMs revolutionalize malware and its development. Their code generation abilities allow malware authors to more easily develop variants, and their logic and command generation capabilities can be used to translate high-level strategic goals into specific terminal commands. Most notably, however, LLMs can be used to turn previously deterministic, scripted malware into malicious agents, with the LLM assuming the role of the "human hacker".

In this semester project, the student will delve into the world of agentic malware powered by an AI assisstant. Specifically, the tasks of the student include:
- Getting familiar with relevant literature
- Develop a Proof-of-Concept implementation for an AI-powered worm