CVE-2026-73325
Received Received - Intake

Unsafe Deserialization in Fujitsu OneCompression Library

Publication date: 2026-08-12

Last updated on: 2026-08-12

Assigner: VulnCheck

Description
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.load_quantized_model_pt() unconditionally calls torch.load with weights_only=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.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-08-12
Last Modified
2026-08-12
Generated
2026-08-12
AI Q&A
2026-08-12
EPSS Evaluated
N/A
NVD
CVE-2026-73325
EUVD
EUVD-2026-57313

Affected Vendors & Products

Showing 1 associated CPE
Vendor Product Version / Range
fujitsu onecompression to 1.2.1 (inc)

Helpful Resources

Exploitability

CWE
CWE Icon
KEV
KEV Icon
CWE ID Description
CWE-502 The product deserializes untrusted data without sufficiently ensuring that the resulting data will be valid.

Attack-Flow Graph

AI Quick Actions

Instant insights powered by AI
Executive Summary

Fujitsu OneCompression library 1.2.0 has an unsafe deserialization flaw in the QuantizedModelLoader.load_quantized_model_pt() function. It calls torch.load with weights_only=False, enabling Python's pickle machinery. Attackers can exploit this by providing a malicious model checkpoint file containing a crafted __reduce__ method to execute arbitrary code when the library loads the file.

Detection Guidance

Check for the presence of Fujitsu OneCompression version 1.2.0 by inspecting installed packages or library files. Look for model.pt files in directories accessible to the application. Monitor for unexpected system commands or processes spawned by the OneCompression library.

Impact Analysis

This vulnerability allows attackers to execute arbitrary code on your system, including system commands, if you load a malicious model.pt file. It can lead to full system compromise, data theft, or further network infiltration depending on the attacker's goals.

Mitigation Strategies

Upgrade Fujitsu OneCompression to version 1.2.1 or later. Restrict write access to directories where model.pt files are loaded. Avoid loading untrusted model files. Implement input validation for model files.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-73325. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart