Goal Reached Thanks to every supporter — we hit 100%!

Goal: 1000 CNY · Raised: 1359 CNY

100%
Buy Us a Coffee

CVE-2026-34445— ONNX: Malicious ONNX models can crash servers by exploiting unprotected object settings.

Quick assessment

Affected
onnx onnx
Exploitation
Public or AI PoC available; prioritize validation
Recommended action
Check the vendor advisory and references for a fixed version. If immediate upgrade is impossible, restrict exposure and increase monitoring.

Open Neural Network Exchange是Open Neural Network Exchange开源的一个开放的生态系统，它使 AI 开发人员能够随着项目的发展选择合适的工具。 Open Neural Network Exchange 1.21.0之前版本存在安全漏洞，该漏洞源于ONNX中的ExternalDataInfo类使用Python的setattr()函数从ONNX模型文件直接加载元数据，但未检查文件中的键是否有效，可能导致攻击者通过特制恶意模型覆盖内部对象属性。

CVSS 8.6 · High EPSS 0.51% · P41

Possible ATT&CK Techniques 1 AI

T1190 · Exploit Public-Facing Application
Get alerts for future matching vulnerabilities Log in to subscribe

I. Basic Information for CVE-2026-34445

Vulnerability Information
Have questions about the vulnerability? See if Shenlong's analysis helps!
View Shenlong Deep Dive ↗

Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.

Vulnerability Title
ONNX: Malicious ONNX models can crash servers by exploiting unprotected object settings.
Source: CVE Program / CVE List V5
Vulnerability Description
Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. Prior to version 1.21.0, the ExternalDataInfo class in ONNX was using Python’s setattr() function to load metadata (like file paths or data lengths) directly from an ONNX model file. It didn’t check if the "keys" in the file were valid. Due to this, an attacker could craft a malicious model that overwrites internal object properties. This issue has been patched in version 1.21.0.
Source: CVE Program / CVE List V5
CVSS Information
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:H
Source: CVE Program / CVE List V5
Vulnerability Type
输入验证不恰当
Source: CVE Program / CVE List V5
Vulnerability Title
Open Neural Network Exchange 安全漏洞
Source: CNNVD (China National Vulnerability Database)
Vulnerability Description
Open Neural Network Exchange是Open Neural Network Exchange开源的一个开放的生态系统，它使 AI 开发人员能够随着项目的发展选择合适的工具。 Open Neural Network Exchange 1.21.0之前版本存在安全漏洞，该漏洞源于ONNX中的ExternalDataInfo类使用Python的setattr()函数从ONNX模型文件直接加载元数据，但未检查文件中的键是否有效，可能导致攻击者通过特制恶意模型覆盖内部对象属性。
Source: CNNVD (China National Vulnerability Database)
CVSS Information
N/A
Source: CNNVD (China National Vulnerability Database)
Vulnerability Type
N/A
Source: CNNVD (China National Vulnerability Database)

Affected Products

Vendor Product Affected Versions CPE Subscribe
onnx onnx < 1.21.0 -

II. Public POCs for CVE-2026-34445

# POC Description Source Link Shenlong Link
AI-Generated POC Verified env Premium
Reproduced successfully in a real sandbox · Below is the actual recording of building the environment and exploiting the vulnerability.
Success marker: VULNERABLE: forged external_data key "basepath" (arbitrary, unvalidated) overwrote internal attribute; attacker value PROOF_62837adfc1aaba75 now in ExternalDataInfo state
Reproduction recording is a Pro+ exclusive
Watch the full sandbox build + live exploit recording for this CVE. Limited-time ¥499/mo.
Upgrade to Pro+
claude_code · 7630 chars
GitHub 复现环境（可 clone 直接跑）
Pro+ exclusive includes:
Vulnerability reproduction recording (real sandbox build + trigger, exclusive)
In-depth vulnerability mechanism
Trigger conditions & impact
Full executable POC code
Exploit chain & mitigation
POC zip download
100+ AI POC generations per month

III. Intelligence Information for CVE-2026-34445

请登录查看更多情报信息。

Patches & Fixes for CVE-2026-34445 (2)

Same Patch Batch · onnx · 2026-04-01 · 4 CVEs total

CVE-2026-34447 5.5 MEDIUM ONNX: External Data Symlink Traversal
CVE-2026-34446 4.7 MEDIUM ONNX: Arbitrary File Read via ExternalData Hardlink Bypass in ONNX load
CVE-2026-27489 ONNX: Path Traversal via Symlink

IV. Related Vulnerabilities

Same product: onnx
Same vendor: onnx
Same weakness: CWE-20

V. Comments for CVE-2026-34445

No comments yet

Leave a comment