CVE-2026-73325: Fujitsu OneCompression 1.2.0 Arbitrary Code Execution via torch.load Deserialization
Fujitsu Research's OneCompression library 1.2.0 contains an unsafe deserialization vulnerability that allows attackers to execute arbitrary code by supplying a crafted model.pt checkpoint file, as QuantizedModelLoader.loadquantizedmodelpt() unconditionally calls torch.load with weightsonly=False, invoking Python's pickle machinery during deserialization. Attackers can embed malicious reduce methods in a crafted model checkpoint to execute arbitrary Python code, including system commands, when the library loads the file from a caller-selected model directory.
Affected Software
Event History
Frequently Asked Questions
What is the severity of CVE-2026-73325?
The severity of CVE-2026-73325 is rated as high with a score of 7.8.
What type of vulnerability is CVE-2026-73325?
CVE-2026-73325 is an arbitrary code execution vulnerability caused by unsafe deserialization in Fujitsu OneCompression.
How do I fix CVE-2026-73325?
To fix CVE-2026-73325, upgrade to Fujitsu OneCompression version 1.2.1 or later where the vulnerability has been addressed.
Which software is affected by CVE-2026-73325?
CVE-2026-73325 affects Fujitsu Research's OneCompression library version 1.2.0.
What can attackers do with CVE-2026-73325?
Attackers can exploit CVE-2026-73325 to execute arbitrary code by supplying a maliciously crafted model.pt checkpoint file.