Explainability study for PDF malware detection, with a public project report. Not production security infrastructure.
-
Updated
Sep 16, 2026
Explainability study for PDF malware detection, with a public project report. Not production security infrastructure.
Malware classification project using static analysis, where applications are converted into bytecode, the byte sequences are transformed into grayscale images, and deep learning–based image classification is applied to categorize malware into 31 distinct subclasses for accurate detection without executing the files.
Local-first phishing URL risk detection with lexical-v4 XGBoost, domain-grouped evaluation, calibrated ALLOW/CAUTION/BLOCK policy, FastAPI, and Chrome extension.
To associate your repository with the security-ml topic, visit your repo's landing page and select "manage topics."