CVE-2026-134172
Received Received - Intake

Insecure Deserialization in Audio-Separator via Malicious Model Files

Vulnerability report for CVE-2026-134172, including description, CVSS score, EPSS score, affected products, exploitability, helpful resources, and attack-flow context.

Publication date: 2026-11-20

Last updated on: 2026-11-20

Assigner: VulnCheck

Description

audio-separator before 0.47.2 contains an insecure deserialization vulnerability that allows attackers to execute code by supplying malicious Demucs .th model files loaded via torch.load with weights_only=False. Local users writing to the shared /tmp/audio-separator-models/ default directory, or attackers controlling unverified model download sources, can execute code when victims run Demucs separation.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-11-20
Last Modified
2026-11-20
Generated
2026-11-21
AI Q&A
2026-11-21
EPSS Evaluated
N/A
NVD
EUVD

Affected Vendors & Products

Showing 1 associated CPE
Vendor Product Version / Range
nomadkaraoke audio-separator From 0 (inc) to 0.47.2 (exc)

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

audio-separator before 0.47.2 has an insecure deserialization flaw. Attackers can exploit this by providing malicious Demucs .th model files. These files are loaded via torch.load with weights_only=False, enabling code execution. The vulnerability exists because local users or attackers can write to the shared /tmp/audio-separator-models/ directory or control unverified model download sources.

Detection Guidance

Detecting this vulnerability requires checking for the presence of audio-separator versions prior to 0.47.2 and verifying if malicious model files exist in the default /tmp/audio-separator-models/ directory. Inspect the installed version of audio-separator using package managers like pip or apt. Search for unexpected or modified .th model files in the shared directory.

Impact Analysis

If you use audio-separator before 0.47.2, an attacker could execute arbitrary code on your system. This could lead to data theft, system compromise, or further network intrusions. The attack requires local access or control over model sources, but the impact is severe due to arbitrary code execution.

Compliance Impact

This vulnerability could lead to unauthorized code execution, potentially exposing sensitive data. For GDPR, it may result in data breaches requiring notification. For HIPAA, it could compromise protected health information. Compliance violations may occur if controls to prevent such attacks are not in place.

Mitigation Strategies

Upgrade audio-separator to version 0.47.2 or later immediately. Remove write permissions from the /tmp/audio-separator-models/ directory to prevent unauthorized file creation. Avoid downloading model files from unverified sources. Monitor the directory for suspicious activity or unknown files.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-134172. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart

Security Radar

News mentioning this vulnerability
Loading related security news...