Is CVE-2026-109375 exploited in the wild?

No. Junglewise Threat Intelligence has no report of CVE-2026-109375 being exploited in the wild.

How severe is CVE-2026-109375?

Medium severity, with a CVSS score of 6.1, published 12 October 2026.

Which products and versions are affected?

Vercel Deepsec 0.1.0 before 1.1.17.

Is there a fix for CVE-2026-109375?

Yes: fixed in 1.1.17. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.

How is CVE-2026-109375 attacked?

It is attacked with local access, in a software library.

Am I affected by CVE-2026-109375?

Check whether you run Vercel Deepsec at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

CVE-2026-109375: Vercel deepsec path traversal in extractTarballLocally

CVE-2026-109375 · Severity: medium · CVSS 6.1 · Published 2026-10-12

Vercel deepsec is a security testing tool that uses coding agents to find vulnerabilities. A flaw in how it extracts tar files allows malicious code running in the sandbox to plant symbolic or hard links that write arbitrary files to the orchestrator host, potentially enabling attackers to escape the sandbox and compromise the system.

Vendors: Vercel.

Quick answers

Is CVE-2026-109375 exploited in the wild?: No. Junglewise Threat Intelligence has no report of CVE-2026-109375 being exploited in the wild.
How severe is CVE-2026-109375?: Medium severity, with a CVSS score of 6.1, published 12 October 2026.
Which products and versions are affected?: Vercel Deepsec 0.1.0 before 1.1.17.
Is there a fix for CVE-2026-109375?: Yes: fixed in 1.1.17. Junglewise finds every place you run it, drafts the upgrade and applies it once you approve.
How is CVE-2026-109375 attacked?: It is attacked with local access, in a software library.
Am I affected by CVE-2026-109375?: Check whether you run Vercel Deepsec at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

How to fix

Fixed in: 1.1.17

Technical details

The extractTarballLocally function in deepsec does not properly validate links within tar archives, allowing sandbox code to plant symbolic or hard links that point outside the extraction directory. An attacker controlling code executed in the sandbox (such as through prompt injection) can exploit this to write arbitrary files on the orchestrator host with the privileges of the deepsec process.

Affected products

Vercel deepsec 0.1.0 before 1.1.17

Timeline

2026-10-12: disclosed

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.