UC Riverside group (Darki, Rokon, Faloutsos, Islam and Papalexakis) have developed a tool called "SourceFinder" that pinpoints online malware source code repositories, such as GitHub, with 89% accuracy. They employed a supervised learning strategy to scan 97,000 malware-related software repositories, locating more than 7,500 malware source code repositories.

The work has been presented at the International Symposium on Research in Attacks, Intrusions, and Defenses (RAID 2020). The research has attracted the attention of security researchers, popular press, security forums (Wilders Security), and is featured at the Communication of the ACM Tech News.

http://hackerchatter.org/sourcefinder/

https://www.wilderssecurity.com/threads/sourcefinder-software-stalks-malware-in-the-wild.434313/

https://cacm.acm.org/news/248924-software-stalks-malware-in-the-wild/fulltext