Skip to content
arXiv· arXiv:2403.04419· Md Rayhanul Masud, Michalis Faloutsos·Indexed by this site · Originally published

Revealing hidden risks: exposing malicious repositories on GitHub under the guise of education

Unveiling A Hidden Risk: Exposing Educational but Malicious Repositories in GitHub

AI summary

A study used ChatGPT to label 35.2K repositories on GitHub that claimed to be for educational purposes only for maliciousness, of which 9294 were judged malicious. Further categorization found 14 malware families, including DDoS, keyloggers, ransomware, etc., and the annual number of releases of such repositories shows an upward trend. The authors say this exploratory study sounds an alarm for the community to better understand and analyze software platforms.

Read originalarxiv.org