CVE-2026-15679
HighCVSS 7.8Exploitation Probability (EPSS)
Low risk38th percentile - higher than 38% of all known CVEs
Summary
Vulnerability in Hugging Face PyTorch Image Models during checkpoint parsing is due to lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. A remote attacker can exploit this flaw to execute code in the context of the current process, requiring user interaction.
Risk Assessment
The risk for the organization is the possibility of remote code execution by an attacker if a user visits a malicious page or opens a malicious file, potentially leading to system compromise.
Recommendation
It is recommended to update Hugging Face PyTorch Image Models to the latest version and educate users to only open trusted files and pages.
Related vulnerabilities
- CVE-2026-120344Critical
A vulnerability was found in the WAVLINK WN551X3-A device in the system function of /bin/ioos, within the WireGuard CLI Group Handler component. Manipulating the group_name argument leads to OS command injection, and the attack can be performed remotely. The exploit has been publicly disclosed and may be used.
- CVE-2026-120322Critical
A vulnerability was identified in Termix-SSH Termix up to 2.9.2 affecting the /plugin-api/file-manager/transferToHost endpoint. Manipulation of the destPath argument leads to path traversal, and the attack can be launched remotely. The exploit is publicly available and may be used.
- CVE-2026-120283Critical
A vulnerability was identified in Linksys E8450 version 1.2.00.360516 in the strcpy function of /www/cgi-bin/portal.cgi. Manipulation of the hidden_timezone argument leads to a stack-based buffer overflow. The attack can be executed remotely, and the vendor did not respond to the disclosure.
- CVE-2026-120282Critical
A vulnerability was determined in WAVLINK WN536AX6-A version V260507 in the host_black function of /protocol.csp. Manipulation of the mac argument causes OS command injection. The exploit has been publicly disclosed and remote exploitation is possible.
- CVE-2026-120281Critical
A vulnerability was found in WAVLINK WN536AX6-A version V260507 in the OpenVPN Handler component, in the strcpy function of /protocol.csp. Manipulation of the openvpn_ip argument results in a stack-based buffer overflow. The exploit has been made public and the attack may be launched remotely.
- CVE-2026-120280Critical
A vulnerability has been found in WAVLINK WN536AX6-A version V260507 in the WireGuard component, in the wireguard_cli_group handler of /protocol.csp. Manipulation of the group_id/group_name arguments leads to OS command injection. The exploit has been disclosed publicly and the attack may be initiated remotely.
- CVE-2026-120279Critical
A flaw has been found in WAVLINK WN536AX6-A version V260507 in the openvpn_cli Handler component, in an unknown function of /protocol.csp. Manipulation of the file_id argument can lead to OS command injection. The exploit has been published and the attack can be launched remotely.
- CVE-2026-120277Critical
A vulnerability was detected in WAVLINK WN536AX6-A (version V260507_2.0.0) in the sub_41a1d8 function of the /bin/ioos file, related to the WireGuard CLI group handler. Manipulation of the group_name argument leads to OS command injection. The issue can be exploited remotely and a public exploit is already available.
- CVE-2026-120276Critical
A weakness was identified in WAVLINK WN536AX6-A (version V260507_2.0.0) in the sub_416760 function of the /bin/ioos file, within the web management interface component. Manipulation of the group_name argument causes OS command injection. The attack can be initiated remotely and the exploit has been made publicly available.
- CVE-2026-120240Critical
A vulnerability was found in the WAVLINK WN553X1-B device (version V260403_2.0.0) in the sub_41A748 function of /bin/ioos, part of the WireGuard CLI Group Handler. Manipulating the group_id/group_name arguments can lead to OS command injection. The attack can be performed remotely and the exploit has been publicly disclosed.
Original NVD description (English source)
Hugging Face PyTorch Image Models checkpoint Deserialization of Untrusted Data Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of Hugging Face PyTorch Image Models. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file. The specific flaw exists within the parsing of checkpoints. The issue results from the lack of proper validation of user-supplied data, which can result in deserialization of untrusted data. An attacker can leverage this vulnerability to execute code in the context of the current process. Was ZDI-CAN-27987.
Vulnerability data from NVD (NIST) · CISA KEV · EPSS