CVE-2025-43849
Published: 05 May 2025
Summary
CVE-2025-43849 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 12.9% of CVEs by exploit likelihood; it is not currently listed in the CISA KEV catalog.
Deeper analysis
Retrieval-based-Voice-Conversion-WebUI is a voice conversion framework based on VITS. Versions 2.2.231006 and earlier contain an unsafe deserialization vulnerability in the model merging workflow. User-supplied paths provided via the ckpt_a and ckpt_b parameters are passed directly to the merge function in process_ckpt.py, which invokes torch.load without safeguards, enabling arbitrary code execution during deserialization.
An unauthenticated remote attacker can supply malicious model paths over the network and trigger the merge operation, achieving remote code execution with the privileges of the web application process. The flaw maps to CWE-502 and carries a CVSS 4.0 score of 8.9, reflecting the absence of required authentication or user interaction.
No patches were available at the time of disclosure. The referenced GitHub Security Lab advisory (GHSL-2025-012) and linked source files document the vulnerable code paths in infer-web.py and process_ckpt.py, indicating that input validation or safe loading mechanisms must be added by operators until an official fix is released.
The EPSS score rose from a low baseline to a peak of 0.0627 on 2026-03-01 before receding to 0.0311, indicating that exploitation interest increased after public disclosure.
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2025-13501
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 ckpt_a and cpkt_b variables take user input (e.g. a path to a model) and pass it to the merge function in…
more
process_ckpt.py, which uses them to load the models on those paths 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.