GitLab Path Traversal Flaw Actively Exploited; CVSS 10.0 Severity
A critical vulnerability in GitLab's commits API allows unauthenticated attackers to read arbitrary files from self-managed servers.
A critical path traversal vulnerability in GitLab is being actively exploited, granting unauthenticated attackers the ability to read arbitrary files from affected servers. The flaw, designated as CVE-2026-85706, carries the maximum possible CVSS severity score of 10.0.
The vulnerability targets the repository commits API, specifically the `/api/v4/projects/{id}/repository/commits/` endpoint. Attackers can exploit the flaw by sending a single HTTP POST request utilizing a `file.path` parameter to bypass security boundaries and access sensitive system files. The vulnerability affects self-managed instances of both GitLab Community Edition (CE) and Enterprise Edition (EE).
GitLab released patches on September 10, 2026—specifically versions 19.3.2, 19.2.6, and 19.1.8. Despite the availability of these updates, the Cybersecurity and Infrastructure Security Agency (CISA) has already added the flaw to its Known Exploited Vulnerabilities (KEV) catalog following confirmed attacks.
A Pattern of Critical Flaws
This security breach was part of a broader update on September 10 that addressed 17 vulnerabilities. Among the other high-risk issues patched in the same release were CVE-2026-87719, an insecure deserialization flaw, and CVE-2026-88765, a buffer overflow vulnerability. The recurrence of high-severity path traversal issues in GitLab's infrastructure highlights a persistent challenge in securing the platform's API endpoints against unauthorized file system access.
Supply Chain Implications
Because GitLab serves as the central backbone for enterprise CI/CD pipelines, the impact of this flaw extends beyond simple data theft. Unauthorized access to the server allows attackers to steal proprietary source code, CI/CD secrets, and administrative credentials. According to Jake Knott, head of threat intelligence at watchTowr, such access enables attackers to inject malicious code into build pipelines, potentially poisoning every downstream environment.
This creates a massive supply-chain risk. By compromising the pipeline, attackers can move laterally into production environments or distribute compromised software to end users. This risk is further compounded by the increasing use of AI coding agents, which often inherit the environment credentials of the systems they operate within, providing another potential vector for credential theft.
Immediate Remediation
Organizations running self-managed GitLab instances must update to the patched versions immediately. Security teams should audit their logs for unusual POST requests to the commits API endpoint. As CISA has confirmed active exploitation, the window for patching is narrow; failure to update leaves the entire software delivery lifecycle vulnerable to total compromise.