criticalZero-Day

GitLab Unauthenticated Path Traversal / Arbitrary File Read (CVE-2026-85706)

First seen Sep 12, 2026 · Updated Sep 12, 2026 · CVSS 10

gitlabpath-traversalunauthenticated-rcefile-readin-the-wild-exploitationsource-code-managementagent-relevant

GitLab disclosed a maximum-severity (CVSS 10.0) path traversal vulnerability in the repository commits API that allows unauthenticated attackers to read arbitrary files on the GitLab server. Active in-the-wild probing was observed within hours of public disclosure, indicating high urgency for patching. Organizations should treat this as an actively exploited zero-day and prioritize immediate remediation.

Technical Analysis

CVE-2026-85706 is a path traversal flaw in GitLab's repository commits API that permits unauthenticated attackers to escape intended directory boundaries and read arbitrary files from the underlying server filesystem, including configuration files, secrets, and potentially SSH/API credentials stored on disk. The CVSS 10.0 rating reflects the lack of authentication requirements combined with full confidentiality impact and network-based exploitability. Threat actors began scanning for and probing vulnerable instances almost immediately after disclosure, a common pattern for high-severity GitLab CVEs given the platform's prevalence in CI/CD and source code management. Exploitation could expose CI/CD secrets, API tokens, environment variables, and internal configuration files used by DevOps pipelines. Many organizations run AI coding agents, CI/CD automation bots, and MCP/tool-use integrations against GitLab instances to fetch repos, trigger pipelines, or manage code; if such agents' API keys, service account tokens, or webhook secrets are stored on the affected GitLab server's filesystem, this flaw could allow attackers to exfiltrate those credentials and pivot into agent-driven automation and downstream systems.

Affected Systems

Self-managed GitLab Community Edition (CE) and Enterprise Edition (EE) instances running vulnerable versions prior to the patched releases; specific affected version ranges should be confirmed against GitLab's official security advisory for CVE-2026-85706. GitLab.com SaaS instances are typically patched centrally by GitLab, but self-hosted deployments remain at risk until manually updated.

Indicators of Compromise

  • No specific hashes, IPs, or domains publicly disclosed at time of reporting; monitor GitLab security advisories and threat intel feeds for updated IOC lists related to CVE-2026-85706 scanning/exploitation activity.
  • Suspicious GET/POST requests to /api/v4/projects/:id/repository/commits or related commits API endpoints with encoded path traversal sequences (e.g., ../, %2e%2e%2f)
  • Unusual file access patterns to /etc/passwd, gitlab.rb, database.yml, or secrets.yml via API-triggered file reads

Remediation Steps

  1. 1

    Patch immediately

    Upgrade all self-managed GitLab instances to the patched versions specified in GitLab's official security release addressing CVE-2026-85706.

  2. 2

    Audit access logs

    Review GitLab API access logs for anomalous requests to the repository commits API, especially those containing path traversal patterns or unexpected file path parameters.

  3. 3

    Rotate exposed secrets

    Rotate GitLab runner tokens, CI/CD variables, SSH keys, API tokens, and any credentials that may have been stored in configuration files potentially exposed by this vulnerability.

  4. 4

    Restrict network exposure

    Limit direct internet exposure of GitLab instances where possible using firewalls, VPNs, or IP allowlisting until patching is confirmed complete.

  5. 5

    Audit agent and automation credentials

    Identify any AI agents, CI/CD bots, or automation tools with API keys or service accounts tied to the affected GitLab instance and rotate those credentials as a precaution.

CVE / Advisory IDs

CVE-2026-85706

Industries Most Exposed

technologysoftware-developmentfinancial-serviceshealthcaregovernmentany-industry-using-gitlab

Sources

Respond to this threat

Pro subscribers get a full AI-generated incident-response playbook for this threat — detection, containment, eradication, and recovery steps — plus an unlimited AI Threat Advisor for questions about your environment.