Automated Detection and Mitigation of Malicious Packages in the PyPI Ecosystem and .exe Files: PyGuardEX

The rapid growth of open-source ecosystems such as PyPI has significantly increased the risk of malicious packages infiltrating and affecting the software supply chains. Attackers often disguise malicious code as legitimate packages, creating severe security threats for end-users. Existing detection methods, mainly signature-based approaches, struggle to detect sophisticated, obfuscated, or dynamically executed malware. To address this growing threat, we propose a machine learning-based tool that scans Python packages and executables to detect malicious behaviours. Our approach leverages, a web-based detection tool PyGuardEX, extracts features, including entropy, and classifies them according to their severity, to provides relevant and detailed information about the malware to prevent future attacks. We present a static, automated malware detection framework that generates real-time alerts, calculates risk scores, and suggests actionable countermeasures. The tool also integrates and displays a constantly updating malware library, providing insights into the latest threats. We validate and test our framework against a wide range of malicious and benign package datasets, and try to achieve high detection accuracy and significantly reduce false positives. Our system can potentially enhance the security of Python's open-source ecosystem, as well as in. exe files providing developers with an efficient tool to mitigate risks posed by malicious packages.