How severe is CVE-2026-90970?

CVE-2026-90970 has a CVSS 3.x base score of 9.9, rated critical severity. It is exploitable over network with low attack complexity, requires low privileges and no user interaction. Impact on confidentiality is high, integrity high, and availability high.

Is CVE-2026-90970 being actively exploited?

It is not currently listed in CISA's KEV catalog. Its EPSS exploit-prediction score is 1% (60th percentile), an estimate of the probability of exploitation in the next 30 days.

How do I fix CVE-2026-90970?

Review the linked vendor and NVD advisories for patched versions and mitigations, then upgrade or apply the recommended workaround. Given its critical severity, prioritise patching exposed systems.

When was CVE-2026-90970 published?

CVE-2026-90970 was published on 2026-10-02.

CVE-2026-90970

CVE-2026-90970 is a critical-severity vulnerability with a CVSS 3.x base score of 9.9. It is not currently listed as actively exploited by CISA, and its EPSS exploit-prediction score is low. The underlying weakness is classified as CWE-1336.

Key facts

Severity: Critical (CVSS 3.x base score 9.9)
EPSS exploit prediction: 1% (60th percentile)
Actively exploited: Not listed in CISA KEV
Weakness: CWE-1336
Published: 2026-10-02
Last modified: 2026-10-02

Description

GitLab has remediated a vulnerability in the GitLab AI Gateway component affecting all versions of the AI Gateway from 18.1.6 before 19.2.4, 19.3 before 19.3.2, and 19.4 before 19.4.1 that, under certain conditions, could have allowed an authenticated user with Duo Agent Platform access to escape the prompt template sandbox via a specially crafted flow configuration, resulting in arbitrary command execution on the AI Gateway.

Frequently asked questions

What is CVE-2026-90970?: GitLab has remediated a vulnerability in the GitLab AI Gateway component affecting all versions of the AI Gateway from 18.1.6 before 19.2.4, 19.3 before 19.3.2, and 19.4 before 19.4.1 that, under certain conditions, could have allowed an authenticated user with Duo Agent Platform access to escape the prompt template sandbox via a specially crafted flow configuration, resulting in arbitrary command execution on the AI Gateway.
How severe is CVE-2026-90970?: CVE-2026-90970 has a CVSS 3.x base score of 9.9, rated critical severity. It is exploitable over network with low attack complexity, requires low privileges and no user interaction. Impact on confidentiality is high, integrity high, and availability high.
Is CVE-2026-90970 being actively exploited?: It is not currently listed in CISA's KEV catalog. Its EPSS exploit-prediction score is 1% (60th percentile), an estimate of the probability of exploitation in the next 30 days.
How do I fix CVE-2026-90970?: Review the linked vendor and NVD advisories for patched versions and mitigations, then upgrade or apply the recommended workaround. Given its critical severity, prioritise patching exposed systems.
When was CVE-2026-90970 published?: CVE-2026-90970 was published on 2026-10-02.

References

https://gitlab.com/gitlab-org/gitlab/-/work_items/628842

Other CWE-1336 vulnerabilities

CVE-2026-97359 — Critical (CVSS 10.0): HFS2 version 2.4.0 and earlier contains a template injection vulnerability in the multipart upload handler that allows…
CVE-2026-75650 — Critical (CVSS 10.0): Adobe Commerce is affected by an Improper Neutralization of Special Elements Used in a Template Engine vulnerability…
CVE-2026-48323 — Critical (CVSS 10.0): Adobe Campaign Classic (ACC) is affected by an Improper Neutralization of Special Elements Used in a Template Engine…
CVE-2026-44181 — Critical (CVSS 10.0): Jupyter Enterprise Gateway launches remote Jupyter Notebook kernels across distributed clusters like Apache Spark,…
CVE-2025-53833 — Critical (CVSS 10.0): LaRecipe is an application that allows users to create documentation with Markdown inside a Laravel app. Versions prior…
CVE-2024-32651 — Critical (CVSS 10.0): changedetection.io is an open source web page change detection, website watcher, restock monitor and notification…

Browse all CWE-1336 vulnerabilities →