CVE-2026-73325

Fujitsu OneCompression < 1.2.1 Arbitrary Code Execution via torch.load Deserialization

Published Aug 12, 2026 CVSS 4.0 · 8.4 High EPSS < 1% KEV · Not listed
Description
Fujitsu Research's OneCompression library before 1.2.1 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.

Analysis

AI-generated

Technical impact

OneCompression library before version 1.2.1 deserializes user‑supplied .pt checkpoint files using torch.load, which triggers Python’s pickle machinery. An attacker can embed a malicious __reduce__ method in a crafted checkpoint that will execute arbitrary Python code, including system commands, when the library loads the file. This untrusted deserialization flaw (CWE-502) enables arbitrary code execution with the privileges of the running process.

Affected Systems

All installations of Fujitsu Research OneCompression 1.2.0 that load external checkpoint files are affected. The issue is fixed in OneCompression 1.2.1, which is available on PyPI.

Risk and Exploitability

The CVSS score of 8.4 indicates high severity and the EPSS score is < 1%, indicating a very low but nonzero exploitation probability. Attackers can exploit the vulnerability by providing a crafted checkpoint file to an application that uses QuantizedModelLoader.load_quantized_model_pt(), which typically requires the ability to supply or write the file to the model directory. Successful exploitation would grant the attacker full control of the application’s operating system context.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Update OneCompression to version 1.2.1 or newer.
  • Restrict the use of external .pt files and validate or whitelist checkpoint files before deserialization.
  • If an update is not immediately possible, isolate the deserialization process in a sandboxed environment or replace torch.load with a safer equivalent that does not trigger Python’s pickle machinery.

Generated by OpenCVE AI on August 25, 2026 at 03:43 UTC.

Tracking

Sign in to view the affected projects.

Advisories

No advisories available
History
Type Values Removed Values Added
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. Fujitsu Research's OneCompression library before 1.2.1 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.
Title Fujitsu OneCompression 1.2.0 Arbitrary Code Execution via torch.load Deserialization Fujitsu OneCompression < 1.2.1 Arbitrary Code Execution via torch.load Deserialization
References
Type Values Removed Values Added
First Time appeared Fujitsu Research
Fujitsu Research onecompression
Vendors & Products Fujitsu Research
Fujitsu Research onecompression
Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'none', 'Technical Impact': 'total'}, 'version': '2.0.3'}

Type Values Removed Values Added
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.
Title Fujitsu OneCompression 1.2.0 Arbitrary Code Execution via torch.load Deserialization
Weaknesses CWE-502
References
Metrics cvssV3_1

{'score': 7.8, 'vector': 'CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H'}

cvssV4_0

{'score': 8.4, 'vector': 'CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N'}

Weaknesses
  • CWE-502

    Deserialization of Untrusted Data

Sources

Status
PUBLISHED
Assigner
VulnCheck
Published
Updated
Aug 25, 2026, 1:15 AM
Reserved
Aug 11, 2026, 9:47 PM
Status
Deferred
Published
Aug 12, 2026, 4:17 PM
Modified
Oct 24, 2026, 3:10 PM