CVE-2026-6855: Red Hat InstructLab path traversal in chat session handler

CVE-2026-6855 · Severity: high · CVSS 7.1 · Published 2026-04-22

InstructLab is an open-source AI model development tool. A security flaw in its chat session handler allows a local user to bypass directory restrictions and write files to unauthorized locations on the system. This could lead to the modification of sensitive system data or the exposure of private information.

Vendors: PyPI, Redhat.

Quick answers

Is CVE-2026-6855 exploited in the wild?
No. Junglewise Threat Intelligence has no report of CVE-2026-6855 being exploited in the wild.
How severe is CVE-2026-6855?
High severity, with a CVSS score of 7.1, published 22 April 2026.
Which products and versions are affected?
instructlab (PyPI) ECOSYSTEM:introduced=0,last_affected=0.26.1.
Is there a fix for CVE-2026-6855?
No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.
How is CVE-2026-6855 attacked?
It is attacked with local access, in a software library.
Am I affected by CVE-2026-6855?
Check whether you run instructlab (PyPI), Redhat Enterprise Linux Ai or Redhat Instructlab 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.

Technical details

A path traversal vulnerability (CWE-22) exists in InstructLab within the chat session handler component. The root cause is the lack of sanitization of the `logs_dir` parameter, which is passed directly to `os.makedirs()` and used in file path construction without boundary validation in `src/instructlab/model/chat.py`. A local attacker with low privileges can exploit this by manipulating the parameter to escape the intended logging directory. This allows for arbitrary directory creation and file writes, potentially leading to unauthorized data modification or disclosure. At the time of reporting, the vulnerability was identified in Red Hat Enterprise Linux AI 3.0 and the InstructLab project.

Affected products

  • Red Hat InstructLab unspecified
  • Red Hat Enterprise Linux AI 3.0

Identifiers

  • CVE-2026-6855
  • GHSA-pqmg-c2j8-fq92
  • PYSEC-2026-2521

Timeline

  • 2026-04-21: disclosed: Initial report in Red Hat Bugzilla
  • 2026-04-22: advisory: CVE published by Red Hat

References

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