By Hamza Faisal
GitLab has rushed out emergency patches for a maximum-severity security flaw that lets an attacker run code on a self-managed GitLab server without logging in at all. The bug, tracked as CVE-2026-110264, carries the highest possible severity score, 10.0 out of 10, and affects every version of the software going back to 2014.
The company released the fixes on October 21 across seven versions at once: 19.5.1, 19.4.3, 19.3.6, 19.2.9, 19.1.9, 19.0.10 and 18.11.13. GitLab says that, under certain conditions, the flaw could allow an unauthenticated user to execute arbitrary code on the server because of an integer overflow when it compiles a specially crafted regular expression. The bug sits in the Onigmo regular expression engine, and researchers say a crafted Org-mode file in a public project, rendered by the server, can trigger it.
A flaw hiding in 12 years of releases
What makes this patch unusual is its reach. Every self-managed GitLab Community Edition and Enterprise Edition release from version 7.0, shipped in 2014, is affected. GitLab does not normally backport fixes to branches this old, yet it pushed the repair all the way back to 18.11, a sign of how seriously it rates the flaw. Security analysts noted that the rare backport alone tells administrators this is not a routine update.
The flaw was found by security researcher Johan Carlsson, known as joaxcar, who reported it through GitLab's HackerOne bug bounty programme. The scoring reflects the worst possible combination for defenders: the attack works over the network, needs no login, needs no user interaction, and its impact extends beyond GitLab itself to connected systems. A successful attack could expose source code, continuous integration secrets, access tokens and connections to production environments. More than half of the Fortune 100 use GitLab's DevSecOps platform, according to the company.
Patching once this month is not enough
Administrators who already patched this month need to patch again. GitLab fixed a different maximum-severity integer-overflow flaw in the same regular expression compiler on October 12, tracked as CVE-2026-97229, and that update does not cover the new bug. Any instance updated to the October 12 versions is back in range for CVE-2026-110264, since the new fixed versions sit one patch above them.
GitLab has also published two detection rules to help administrators hunt for exploitation attempts in their logs, named Onigmo native worker crash and Org mode render error burst. The company says it has seen no sign the flaw has been used in attacks, and the bug is not listed in the United States CISA catalogue of known exploited vulnerabilities as of October 22. That reassurance has a time limit, though: now that the patch is public, attackers can study exactly what changed.
Who needs to act, and who does not
Only self-managed installations need to act. GitLab.com already runs the fixed version, and GitLab Dedicated tenants on the internal releases 19.3.5-internal0 or 19.4.2-internal0 and later already have the fix. Everyone else should upgrade to the fixed version for their release line, starting with internet-facing servers.
The same release also fixed eleven lower-severity issues, including a high-severity LDAP account-takeover flaw and an OAuth and OpenID Connect cache-poisoning bug. And the underlying engine problem runs wider than GitLab: Ruby shipped the Onigmo fixes behind this CVE and three related ones in versions 4.0.9, 3.4.12 and 3.3.13 on October 23. For teams running their own GitLab servers, the message is blunt: this is the second perfect-score flaw in the same component in nine days, and the patch window is now.


