hn.today

26,400 live credentials found scanning 7.25M repos on China's Gitee

mongodb.com28 points6 comments
Screenshot of 26,400 live credentials found scanning 7.25M repos on China's Gitee

A large-scale credential-exposure investigation scanned 7,256,531 public repositories on Gitee, China’s largest domestic Git-hosting platform. Because Gitee’s official API offers limited access and its code search is constrained, the researcher pieced together a repository list from user searches, Explore feeds, users’ project and starred-repository pages, and the internal search service powering Gitee’s website. The process built a corpus of 1,704,759 usernames and used recursive, subdivided searches to find repositories beyond that social network. The work highlights how secrets can remain exposed in Git history and how limited platform-wide scanning leaves risks difficult to identify.

Repositories were bare-cloned using a serverless compute setup, then scanned concurrently with TruffleHog and MongoDB’s Kingfisher. Both tools checked candidate secrets against vendor services, so only credentials that authenticated at scan time were counted; duplicates were removed by secret value. The combined scan found 26,400 unique live credentials: TruffleHog found 12,737, Kingfisher 14,664, and 1,001 were detected by both. Findings included thousands of active cloud-provider access keys and an organization-admin credential for a major enterprise AI platform. Running the scan across 350 concurrent workers took about a month and cost $3,097.64, covered by compute-credit sponsorship. The researcher argues that the same dual-scanner approach can be adapted to other Git-hosting services.

Read on mongodb.com6 comments on Hacker News

Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.

More in Security

The daily digest

Today's best Hacker News stories, summarized and screenshotted, one email a day.