Research · curated 7 Aug 2026
Defending FCG-based Malware Detectors Against Metamorphic Android Variants via LLM-Powered Code Refactoring
First reported openreview.net
Coverage timeline
Single-source research — first reported, latest, and curated coincide.
Why it matters
FCGA addresses an emerging attack surface where adversaries use LLMs to rewrite malicious code and evade static malware classifiers, a defensive concern for anyone deploying ML-based malware detection.
A research paper (ACL ARR 2026 submission) introduces FCGA, a framework that uses LLMs to synthesize code augmentations for training function-call-graph-based Android malware classifiers, hardening them against LLM-rewritten metamorphic malware variants. The authors report that graph-based detectors enriched with LLM features drop over 60% in accuracy against LLM-rewrite attacks, and FCGA improves robustness by up to 8% over baselines.