ALIBI: Adversarial Legitimacy Injection in Binary Input against LLM Malware Analyzers

Factual evidence

What the source reports

Researchers reveal ALIBI, an attack inserting false read-only text into binaries to trick LLM-based malware analyzers.