CVE-2025-43852
Published: 05 May 2025
Summary
CVE-2025-43852 is a high-severity Deserialization of Untrusted Data (CWE-502) vulnerability in Rvc-Project Retrieval-Based-Voice-Conversion-Webui. Its CVSS base score is 8.9 (High).
Operationally, ranked in the top 12.5% of CVEs by exploit likelihood; it is not currently listed in the CISA KEV catalog.
Deeper analysis
Retrieval-based-Voice-Conversion-WebUI versions 2.2.231006 and prior are affected by an unsafe deserialization flaw in the VITS-based voice conversion framework. The model_choose variable accepts a user-supplied path that is forwarded to the uvr function in vr.py; when the name contains the substring "DeEcho", an AudioPreDeEcho instance is created whose model_path attribute is later passed directly to torch.load, allowing a malicious serialized object to execute arbitrary code during deserialization.
An unauthenticated remote attacker can supply a crafted path through the web interface to achieve remote code execution, resulting in full compromise of confidentiality, integrity, and availability on the host. The CVSS 4.0 score of 8.9 reflects the network attack vector, low complexity, and absence of required privileges or user interaction.
The GitHub Security Lab advisory GHSL-2025-012_GHSL-2025-022_Retrieval-based-Voice-Conversion-WebUI and the linked source references confirm that no patches existed at publication time.
The issue occurs in an AI/ML voice-conversion application that relies on PyTorch model loading. EPSS rose from lower values to a peak of 0.0602 on 2026-03-01 before receding, indicating that exploitation interest increased after disclosure.
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2025-13504
Vulnerability details
Retrieval-based-Voice-Conversion-WebUI is a voice changing framework based on VITS. Versions 2.2.231006 and prior are vulnerable to unsafe deserialization. The model_choose variable takes user input (e.g. a path to a model) and passes it to the uvr function in vr.py. In…
more
uvr , if model_name contains the string "DeEcho", a new instance of AudioPreDeEcho class is created with the model_path attribute containing the aforementioned user input. In the AudioPreDeEcho class, the user input is used to load the model on that path with torch.load, which can lead to unsafe deserialization and remote code execution. As of time of publication, no known patches exist.
- CWE(s)
Related Threats
No named actor attribution yet. ATT&CK technique mapping in progress for this CVE.
Affected Assets
Mitigating Controls
Likely Mitigating Controls AI
Per-CVE control mapping for this CVE has not run yet; the list below is derived from the weakness types (CWEs) cited in the NVD entry.
Penetration testing supplies malicious serialized objects, detecting unsafe deserialization and supporting corrective actions.
Evaluation of untrusted data handling (deserialization testing) reveals unsafe processing, which the required remediation process addresses.
Untrusted serialized data can be deserialized and observed inside the chamber, blocking gadget-chain exploitation outside the sandbox.
Validates or rejects untrusted serialized data before deserialization occurs.
Identifies and blocks malicious code introduced through deserialization of untrusted data at system boundaries.
Integrity verification of serialized information can detect tampering before deserialization occurs.
Provenance of associated data allows detection of untrusted sources before deserialization or processing occurs.