CVE-2026-79785
Model downloader disables TLS, enabling MITM-supplied models and remote code execution.
Is CVE-2026-79785 being exploited?
Not confirmed. CVE-2026-79785 does not appear in CISA's Known Exploited Vulnerabilities catalog, which records only exploitation that has been observed and reported publicly. That is evidence of absence of a report, not evidence the flaw is unattacked. EPSS currently estimates a 0.18% probability of exploitation in the next 30 days.
How severe is CVE-2026-79785?
CVE-2026-79785 is rated High. Severity describes how bad exploitation would be, not how likely it is: pair it with exploitation evidence before deciding what to patch first.
Is there a patch for CVE-2026-79785?
Not known from our data. No patch reference has been recorded for CVE-2026-79785, which is not the same as no patch existing — a fix may have shipped without a tagged reference, or after this record was written. The vendor advisory is the only authority on whether a fix exists, and mitigations may be available regardless.
What does CVE-2026-79785 affect?
CVE-2026-79785 affects X-AnyLabeling (anylabeling/services/auto_labeling/model.py), Deployments auto-downloading models from project release host, PyTorch .pth/.pt workflows (SAM2 video, YOLOE, UPN, open_vision configs), ONNX model consumers using downloaded models. Confirm the exact affected versions against the vendor advisory before deciding you are exposed.
What should I do about CVE-2026-79785?
Disable auto-downloads/block model hosts; validate model hashes/signatures; restrict network access; apply vendor patch when released.
Exploitation status reflects CISA's KEV catalog as we last synced it. Check the catalog directly.
MITM can deliver malicious model: cause remote code execution via PyTorch unpickle or manipulate inference outputs.
Immediate action required
- affected< 4.0.0-beta.9
- affected< 4.0.0b9
As published in the CVE Program record. A version outside these ranges is not a statement that it is unaffected — vendors sometimes understate a range, and distribution-backported builds carry upstream numbers that do not reflect what was patched into them.