A Critical GitLab Flaw Has Been Exploited Just Days After Disclosure, Leaving Users Exposed to Serious Risks
In a shocking turn of events, threat actors have begun exploiting a critical-severity vulnerability in GitLab just two days after it was publicly disclosed. The flaw, tracked as CVE-2026-19478 and boasting a CVSS score of 9.4, allows an unauthenticated attacker to remotely modify or delete public projects and user data via a GraphQL directive. This means that anyone with internet access can potentially wreak havoc on a GitLab instance without needing any login credentials.
The vulnerability was patched on August 17 in various versions of GitLab Community Edition (CE) and Enterprise Edition (EE). However, it appears that some users may have been too slow to update their instances, leaving them vulnerable to exploitation. WatchTowr, an attack surface management company, warned on August 18 that the flaw could be easily reproduced and urged users to update their self-managed instances as soon as possible. They also recommended restricting unauthenticated access to the /api/graphql endpoint or removing public repository access entirely.
The rapid reproduction of this vulnerability by WatchTowr’s researchers is particularly concerning, as it highlights the growing threat posed by AI-enabled attackers. “Whilst no public exploit code is available, we were able to reproduce the vulnerability within minutes of its disclosure, armed only with the advisory details and patch,” a WatchTowr spokesperson said. This means that sophisticated attackers are likely to be hot on the heels of the researchers, using this knowledge to launch targeted attacks.
The use of AI in vulnerability exploitation has become increasingly prevalent, and it’s clear that this is the new reality we’re facing. Jake Knott, principal security researcher at WatchTowr, notes that the narrow patch window for this vulnerability is a major concern. “The newly disclosed code injection vulnerability allows an unauthenticated attacker to delete publicly accessible GitLab projects and rewrite their state, deleting repositories entirely, forging merge records, or banning maintainers in a single HTTP request with no credentials, user interaction, or obscure configuration required,” he explains.
The implications of this vulnerability are far-reaching, particularly when it comes to supply chain attacks. Patrick Münch, co-founder and CSO at Mondoo, warns that the ability to forge any merge records makes it trivial for attackers to insert malicious code into a project without anyone noticing. “Think about what that means,” he says. “In every supply chain worm we’ve tracked this year, attackers had to work around code review because faking an approval was harder than dodging it. This does it for free.”
Given the severity of this vulnerability and the ease with which it can be exploited, it’s essential that users take immediate action to protect their instances. If you haven’t already patched your GitLab instance, do so as soon as possible. Additionally, monitor your web logs for any signs of attempted exploitation or probes containing the string ‘@gl_introduced’. This may help you detect and prevent potential attacks before they’re successful.
The rapid exploitation of this vulnerability serves as a stark reminder of the importance of staying up-to-date with security patches and being vigilant about monitoring for suspicious activity. In today’s threat landscape, complacency is no longer an option – only timely action can ensure that your data remains secure.
Source: SecurityWeek — 2026-08-20