Bug 2459998 (CVE-2026-6859) - CVE-2026-6859 instructlab: InstructLab: Arbitrary code execution due to hardcoded `trust_remote_code=True`
Summary: CVE-2026-6859 instructlab: InstructLab: Arbitrary code execution due to hardc...
Keywords:
Status: NEW
Alias: CVE-2026-6859
Product: Security Response
Classification: Other
Component: vulnerability
Version: unspecified
Hardware: All
OS: Linux
high
high
Target Milestone: ---
Assignee: Product Security
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2026-04-21 07:51 UTC by OSIDB Bzimport
Modified: 2026-04-22 12:59 UTC (History)
1 user (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed:
Embargoed:

Attachments (Terms of Use)

Description OSIDB Bzimport 2026-04-21 07:51:02 UTC 
InstructLab hardcodes trust_remote_code=True in linux_train.py for all HuggingFace from_pretrained() calls. This enables arbitrary Python code execution from malicious model repositories on HuggingFace Hub. Attacker needs only a free HuggingFace account; victim runs ilab train/download/generate with the malicious model name.

Upstream notification: security-reporting bounced for external senders. Filed via Red Hat.
Note You need to log in before you can comment on or make changes to this bug.