26,400 live credentials found scanning 7.25M repos on China's Gitee (www.mongodb.com)

A recent security scan by independent researcher Ben Zimmermann, in collaboration with MongoDB's Kingfisher team, uncovered a staggering 26,400 live credentials while examining 7.25 million public repositories on Gitee, China's primary Git hosting platform. This significant discovery highlights the vulnerability of credential exposure on Gitee, especially given its limited API and lack of established secret-scanning partnerships seen on platforms like GitHub. Notable finds included credentials related to major enterprise AI platforms, emphasizing the scope of the threat. The investigation utilized a dual-scanner pipeline combining TruffleHog and Kingfisher to maximize the identification of exposed credentials, with results indicating that roughly 1 in every 766 Gitee repositories contained a live Alibaba Cloud access key. This finding is instrumental for the AI/ML community as it reveals that major tech players, both Chinese and Western, are contributing to credential leaks, suggesting a widespread issue that transcends geographical boundaries. Notifications were sent to affected vendors, yet the lack of coordinated security channels within many Chinese companies complicates remediation efforts, marking a critical area for future focus in managing and securing sensitive information.