Skip to main content
Vulnerability Database/CVE-2026-92470

CVE-2026-92470: GitLab EE Information Disclosure Vulnerability

CVE-2026-92470 is an information disclosure vulnerability in GitLab Enterprise Edition that exposes sensitive CI/CD variables through Duo AI troubleshooting. This article covers technical details, affected versions, and mitigation.

Updated:

CVE-2026-92470 Overview

CVE-2026-92470 is a missing authorization vulnerability [CWE-862] in GitLab Enterprise Edition (EE). The flaw exists in the Duo AI troubleshooting feature. Under specific conditions, an authenticated user can retrieve sensitive Continuous Integration/Continuous Delivery (CI/CD) variable values from debug-mode job traces. The issue affects GitLab EE versions from 18.7 before 19.2.7, 19.3 before 19.3.3, and 19.4 before 19.4.1. GitLab remediated the flaw in patch release 19.4.1 and the associated backport releases.

Critical Impact

An authenticated attacker can access sensitive CI/CD variables through the Duo AI troubleshooting feature, potentially exposing pipeline secrets, tokens, and credentials.

Affected Products

  • GitLab EE versions 18.7 through 19.2.6
  • GitLab EE version 19.3 before 19.3.3
  • GitLab EE version 19.4 before 19.4.1

Discovery Timeline

  • 2026-09-24 - CVE-2026-92470 published to the National Vulnerability Database (NVD)
  • 2026-09-24 - Last updated in NVD database

Technical Details for CVE-2026-92470

Vulnerability Analysis

The vulnerability resides in the Duo AI troubleshooting workflow inside GitLab EE. When a CI/CD job runs with debug mode enabled, the resulting job trace contains expanded variable values. The Duo AI troubleshooting feature ingests these traces to help users diagnose job failures. Missing authorization checks in this feature allow an authenticated user to request troubleshooting output for jobs they should not be able to inspect. The response can include the sensitive variable values embedded in the debug trace.

Because the attack requires only low privileges and executes over the network without user interaction, the impact scope extends beyond the vulnerable component to any project whose debug traces the feature can reach.

Root Cause

The root cause is a missing authorization check [CWE-862] in the Duo AI troubleshooting request handler. The feature retrieves job trace data without validating that the requesting user has permission to view the underlying job or its variables. Debug-mode traces store variable values in plaintext, so the authorization gap directly translates into secret disclosure.

Attack Vector

An authenticated GitLab user submits a Duo AI troubleshooting request that references a targeted job. If the target job ran with debug tracing enabled and the missing authorization condition is met, the returned troubleshooting content reveals CI/CD variable values from that trace. No social engineering, victim interaction, or elevated role is required beyond having a valid authenticated session.

No public proof-of-concept exploit is currently available. See the GitLab Patch Release 19.4.1 and GitLab Work Item #617719 for vendor-provided technical detail.

Detection Methods for CVE-2026-92470

Indicators of Compromise

  • Unexpected Duo AI troubleshooting API requests referencing jobs across projects the user does not own or maintain.
  • Access log entries showing repeated troubleshooting queries against jobs that used CI_DEBUG_TRACE or CI_DEBUG_SERVICES.
  • Application logs indicating retrieval of job traces by users lacking Developer or higher role on the source project.

Detection Strategies

  • Correlate Duo AI feature usage events with the project membership of the requesting user to surface cross-project access.
  • Alert on any pipeline that enables debug tracing in combination with production-scoped protected variables.
  • Review GitLab audit events for job_artifact or trace access patterns that deviate from a user's normal project scope.

Monitoring Recommendations

  • Enable and forward GitLab application, audit, and production logs to a centralized analytics platform for retention and query.
  • Track the count of Duo AI troubleshooting invocations per user and baseline normal activity to detect anomalies.
  • Monitor CI/CD variable rotation events and any post-incident secret regeneration to confirm remediation coverage.

How to Mitigate CVE-2026-92470

Immediate Actions Required

  • Upgrade GitLab EE to version 19.4.1, 19.3.3, or 19.2.7 depending on the current release branch.
  • Rotate all CI/CD variables that may have been exposed through debug-mode job traces during the vulnerable window.
  • Disable CI_DEBUG_TRACE on production pipelines and restrict its use to isolated debugging environments.
  • Audit Duo AI troubleshooting activity across the instance to identify potentially exposed jobs.

Patch Information

GitLab addressed CVE-2026-92470 in patch release 19.4.1, with backports to 19.3.3 and 19.2.7. Administrators running any affected version should upgrade to the fixed release corresponding to their branch. Full details are available in the GitLab Patch Release 19.4.1 announcement.

Workarounds

  • Disable the Duo AI troubleshooting feature at the instance or group level until the patch is applied.
  • Mark all sensitive CI/CD variables as masked and protected to reduce residual exposure in trace outputs.
  • Restrict who can trigger jobs with debug tracing enabled through project-level pipeline permissions.
bash
# Verify GitLab version and upgrade guidance
sudo gitlab-rake gitlab:env:info | grep -i version

# Example: upgrade an Omnibus GitLab EE instance on Debian/Ubuntu
sudo apt-get update
sudo apt-get install gitlab-ee=19.4.1-ee.0

# Disable debug tracing at the project or group level by removing
# CI_DEBUG_TRACE from CI/CD variables in the GitLab UI or via API.

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Default Legacy - Prefooter | Experience the World’s Most Advanced Cybersecurity Platform

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.