---
id: CVE-2026-6859
aliases:
  - GHSA-rxpq-xgqx-fr7p
  - PYSEC-2026-2522
title: InstructLab Includes Functionality from Untrusted Control Sphere
summary: InstructLab Includes Functionality from Untrusted Control Sphere
severity: high
cvss: 8.8
cvssVector: 'CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H'
vendor: instructlab
product: instructlab
ecosystem: pip
affected:
  - instructlab <= 0.26.1
published: '2026-04-22'
updated: '2026-07-13'
source: OSV
sourceUrl: 'https://osv.dev/vulnerability/GHSA-rxpq-xgqx-fr7p'
references:
  - url: 'https://nvd.nist.gov/vuln/detail/CVE-2026-6859'
  - url: 'https://access.redhat.com/security/cve/CVE-2026-6859'
  - url: 'https://bugzilla.redhat.com/show_bug.cgi?id=2459998'
  - url: 'https://github.com/instructlab/instructlab'
tags:
  - osv
  - pip
epss: 0.00441
epssPercentile: 0.37769
ingestedAt: '2026-07-13T18:58:03.046Z'
---

## Overview

A flaw was found in InstructLab. The `linux_train.py` script hardcodes `trust_remote_code=True` when loading models from HuggingFace. This allows a remote attacker to achieve arbitrary Python code execution by convincing a user to run `ilab train/download/generate` with a specially crafted malicious model from the HuggingFace Hub. This vulnerability can lead to complete system compromise.

## Affected packages

- `instructlab <= 0.26.1`

## Remediation

Refer to the advisory for the patched release.