Loading vulnerability…
CVE-2026-34445 affects Open Neural Network Exchange (ONNX) versions prior to 1.21.0. The flaw is in the ExternalDataInfo class, which used Python's setattr() to populate metadata fields from an ONNX model's external data entries, including values such as file paths and data lengths. The implementation did not validate that attacker-controlled keys from the model file corresponded only to legitimate, expected attributes. As a result, a crafted malicious ONNX model could cause arbitrary internal object properties to be overwritten during model parsing/loading. This is an unsafe dynamic attribute assignment issue that can corrupt object state and alter parser behavior. The issue was patched in ONNX 1.21.0.
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
If you can’t patch tonight, do this now.
Patch, then assume compromise.
No public exploits tracked yet. Mallory keeps watching.
No public exploit code observed for this vulnerability.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
7 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.