Tencent NeuralNLP-NeuralClassifier Untrusted Data Deserialization RCE (CVE-2025-13708)

Key vulnerability information extracted from the web screenshot: - **Vulnerability Title**: - Tencent NeuralNLP-NeuralClassifier _load_ checkpoint Deserialization of Untrusted Data Remote Code Execution Vulnerability - **Vulnerability Identifiers**: - ZDI-25-1033 - ZDI-CAN-27184 - CVE-2025-13708 - **CVSS Score**: - 7.8 (AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H) - **Affected Vendor**: - Tencent - **Affected Product**: - NeuralNLP-NeuralClassifier - **Vulnerability Details**: - This vulnerability allows remote attackers to execute arbitrary code on affected Tencent NeuralNLP-NeuralClassifier installations. Exploitation requires user interaction. - The specific vulnerability resides in the _load_ checkpoint function, where insufficient validation of user-supplied data leads to deserialization of untrusted data. Attackers can exploit this to execute code with root privileges. - **Additional Details**: - Tencent has released an update to fix this vulnerability. More details are available at: - - **Disclosure Timeline**: - 2025-05-22 - Vulnerability reported to vendor - 2025-12-01 - Coordinated vulnerability disclosure - 2025-12-01 - Advisory update - **Acknowledgments**: - Peter Girnus (@gothburz) from Trend Zero Day Initiative

Goal Reached Thanks to every supporter — we hit 100%!

Tencent NeuralNLP-NeuralClassifier Untrusted Data Deserialization RCE (CVE-2025-13708)