CVE-2026-6859 — A flaw was found in InstructLab. The `linux_train.py` script hardcodes `trust_r…
Severity: High · Kind: Vulnerability
We've ingested this item but haven't summarised it yet. Read the upstream advisory using the link below in the meantime — the AI summary will appear here once the next run completes.
From the source
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.
Vulnerability facts
- CVE
- CVE-2026-6859
- CVSS
- 8.8
Was this useful?
Plain-English summaries are AI-generated and reviewed for tone, not technical accuracy. For incident response, always rely on the original source linked above.