Research · curated 19 Sep 2026
Malicious Model Detection with MLTracer Syscalls
First reported binarly.io
Coverage timeline
Single-source research — first reported, latest, and curated coincide.
Why it matters
MLTracer's findings show that widely used static model scanners on Hugging Face miss a large share of malicious pickle-based models, exposing defenders relying on those scanners to AI supply-chain code-execution attacks.
Binarly researchers present MLTracer, a pipeline that detects malicious ML model files through dynamic analysis of system-call traces captured in a sandbox, labeling load-time behavior with syscall categories and a fine-tuned Sentence Transformers embedding model. Applied at scale to Hugging Face, the study found major platform scanners miss over 20% of models that trigger unintended code execution and systematizes the static-scanner evasion techniques behind those misses.