AI-Assisted Malware Analysis with Local Open-Source LLMs

No ratings

Presented at CONFidence 2026 by

Paweł Zegartowski

A practical session on using local open-source / open-weights LLMs for malware analysis: local setup, real limitations, multi-agent workflows with Ghidra, IDA, and x64dbg, and a live Xloader malware analysis showing where these models help and where they fail.