Skip to main content
COSMICBYTEZLABS
NewsSecurityHOWTOsToolsStudyTraining
ProjectsChecklistsAI RankingsNewsletterStatusTagsAbout
Subscribe

Press Enter to search or Esc to close

News
Security
HOWTOs
Tools
Study
Training
Projects
Checklists
AI Rankings
Newsletter
Status
Tags
About
RSS Feed
Reading List
Subscribe

Stay in the Loop

Get the latest security alerts, tutorials, and tech insights delivered to your inbox.

Subscribe NowFree forever. No spam.
COSMICBYTEZLABS

Your trusted source for IT intelligence, cybersecurity insights, and hands-on technical guides.

469+ Articles
115+ Guides

CONTENT

  • Latest News
  • Security Alerts
  • HOWTOs
  • Projects
  • Exam Prep

RESOURCES

  • Search
  • Browse Tags
  • Newsletter Archive
  • Reading List
  • RSS Feed

COMPANY

  • About Us
  • Contact
  • Privacy Policy
  • Terms of Service

© 2026 CosmicBytez Labs. All rights reserved.

System Status: Operational
  1. Home
  2. Security
  3. CVE-2026-2370: GitLab Jira Connect Credential Impersonation
CVE-2026-2370: GitLab Jira Connect Credential Impersonation
SECURITYHIGHCVE-2026-2370

CVE-2026-2370: GitLab Jira Connect Credential Impersonation

GitLab has patched a high-severity vulnerability in its Jira Connect integration affecting CE/EE versions from 14.3 through 18.10 that allowed an authenticated low-privilege user to obtain installation credentials and impersonate other workspace members.

Dylan H.

Security Team

March 30, 2026
5 min read

Affected Products

  • GitLab CE/EE >= 14.3 < 18.8.7
  • GitLab CE/EE >= 18.9 < 18.9.3
  • GitLab CE/EE >= 18.10 < 18.10.1

CVE-2026-2370: GitLab Jira Connect Credential Impersonation

A high-severity vulnerability has been disclosed in GitLab CE/EE, tracked as CVE-2026-2370 with a CVSS score of 8.1 (High). The flaw exists in GitLab's Jira Connect integration and affects all versions from 14.3 through 18.10, with patches available in 18.8.7, 18.9.3, and 18.10.1. An authenticated user with minimal workspace permissions can exploit the vulnerability to obtain Jira Connect installation credentials and use them to impersonate other users within the connected Jira workspace.


Vulnerability Overview

AttributeValue
CVE IDCVE-2026-2370
CVSS Score8.1 (High)
Affected ComponentGitLab Jira Connect Integration
Affected VersionsCE/EE 14.3 – 18.8.6, 18.9.0 – 18.9.2, 18.10.0
Patched Versions18.8.7, 18.9.3, 18.10.1
Attack VectorAuthenticated (minimal permissions required)
In-the-Wild ExploitationNot confirmed at time of disclosure
PublishedMarch 30, 2026

Technical Analysis

The vulnerability resides in the Jira Connect OAuth integration layer within GitLab. The Jira Connect feature allows GitLab projects to link with Atlassian Jira Software installations, enabling cross-referencing of issues, branches, commits, and merge requests between the two platforms.

To facilitate this integration, GitLab generates and stores installation credentials — secrets used to authenticate API calls from GitLab to the connected Jira installation. These credentials represent the bridge between the GitLab project context and the Jira workspace.

The core flaw is that the endpoint responsible for serving or validating Jira Connect installation state does not enforce sufficient authorization checks. An authenticated GitLab user with no more than basic membership permissions (e.g., Guest or Reporter role) can craft requests to retrieve the Jira Connect installation token or signing secret for a connected Jira installation.

Exploitation Path

1. Attacker holds any authenticated GitLab account (Guest-level or higher)
   on a GitLab instance with Jira Connect configured
 
2. Attacker queries the GitLab Jira Connect integration endpoint — no
   elevated permissions are required to access the credential material
 
3. Obtained installation credentials allow the attacker to generate signed
   Jira Connect API requests authenticated as the GitLab integration
 
4. Attacker leverages the credentials to impersonate other users within the
   connected Jira workspace — performing actions under their identity
 
5. Depending on Jira workspace permissions, impersonated actions may include
   reading sensitive issues, modifying tickets, or accessing project data

Scope and Prevalence

GitLab is one of the most widely deployed DevSecOps platforms globally, used across enterprise, government, and open-source communities. The Jira Connect integration is a commonly enabled feature in organizations that use both GitLab and Atlassian Jira for project and issue tracking.

The breadth of the affected version range — from GitLab 14.3 (released in 2021) through current 18.10.0 — means that a substantial portion of GitLab deployments worldwide could be vulnerable if not yet patched. Managed GitLab.com is typically updated promptly, but self-managed instances may lag behind.


Remediation

Primary Fix: Upgrade GitLab

GitLab has issued patched releases addressing CVE-2026-2370:

StreamPatched Version
18.8.x18.8.7
18.9.x18.9.3
18.10.x18.10.1

Update your self-managed GitLab instance to one of these versions immediately.

# For Omnibus installations — check current version
gitlab-rake gitlab:env:info | grep "GitLab version"
 
# Upgrade via package manager (Debian/Ubuntu)
apt-get update && apt-get install gitlab-ee
 
# For GitLab Runner or source installations, follow the standard upgrade guide

Interim Mitigations

If an immediate upgrade is not possible:

  1. Disable Jira Connect — navigate to Admin Area > Settings > Integrations and disable the Jira Connect application until the instance is patched
  2. Audit Jira Connect tokens — rotate any existing Jira Connect installation credentials through the Jira Software administration panel
  3. Review GitLab audit logs — check for unexpected API calls to Jira Connect integration endpoints from low-privilege accounts

Detection

Identify potential exploitation attempts in GitLab logs:

# GitLab Omnibus — check production and API logs for Jira Connect endpoint access
grep -i "jira_connect" /var/log/gitlab/gitlab-rails/production.log | \
  grep -v "200" | tail -100
 
# Review API access log for anomalous Jira Connect requests
grep "jira_connect" /var/log/gitlab/gitlab-rails/api_json.log | \
  python3 -c "import sys, json; [print(json.loads(l).get('username',''), json.loads(l).get('path',''), json.loads(l).get('status','')) for l in sys.stdin]"
 
# Check audit events for integration configuration access
gitlab-rails runner "AuditEvent.where(\"details LIKE '%jira_connect%'\").last(50).each { |e| puts e.details }"

Impact Assessment

Impact AreaDescription
User ImpersonationAttacker can act as other users within the connected Jira workspace
Data AccessAccess to Jira issues, project boards, and linked development data
Privilege ScopeBounded by permissions of the impersonated Jira account
Attack BarrierRequires authenticated GitLab account on an affected instance
Cascade RiskCompromised Jira credentials may affect other Atlassian integrations

Key Takeaways

  1. CVE-2026-2370 is a CVSS 8.1 High flaw in GitLab's Jira Connect integration allowing credential theft and user impersonation
  2. Affects a very wide version range: GitLab CE/EE from 14.3 all the way to 18.10.0
  3. Only an authenticated account with minimal permissions is required — no admin access needed
  4. Upgrade to 18.8.7, 18.9.3, or 18.10.1 immediately; rotate Jira Connect credentials if exploitation cannot be ruled out
  5. GitLab.com (SaaS) is typically patched automatically; self-managed instances require manual action

Sources

  • CVE-2026-2370 — NIST NVD
  • GitLab Security Advisories
#CVE-2026-2370#GitLab#Jira Connect#Credential Theft#NVD#Impersonation

Related Articles

CVE-2025-15036: MLflow Path Traversal in Archive Extraction

A critical path traversal vulnerability in MLflow's extract_archive_to_dir function allows attackers to write arbitrary files outside the intended extraction directory via maliciously crafted tar archives. Affects all versions before v3.7.0.

6 min read

CVE-2025-15379: MLflow Command Injection in Model Serving (CVSS 10.0)

A maximum-severity command injection vulnerability in MLflow's model serving container initialization allows attackers to execute arbitrary OS commands via a maliciously crafted python_env.yaml dependency file when deploying models with env_manager=LOCAL.

7 min read

CVE-2026-32973: OpenClaw Exec Allowlist Bypass via Glob Wildcard Overmatch

A critical CVSS 9.8 vulnerability in OpenClaw allows attackers to bypass the exec allowlist by exploiting improper glob pattern normalization where the ? wildcard ovematches across POSIX path segments.

5 min read
Back to all Security Alerts