CVE-2025-60017 is a high-severity vulnerability affecting Unitree robotic devices including the Go2, G1, H1, and B2 models. The flaw is an OS command injection (CWE-78) occurring in the hostapd_restart.sh script, specifically via the wifi_ssid or wifi_pass parameters used in the restart_wifi_ap and restart_wifi_sta functions. These parameters are not properly sanitized, allowing an attacker to inject arbitrary OS commands that execute with root privileges. The vulnerability is remotely exploitable (Attack Vector: Adjacent Network) but requires high attack complexity and no user interaction. The scope is changed (S:C), meaning the vulnerability can affect resources beyond the initially vulnerable component. The CVSS 3.1 base score is 8.2, reflecting low confidentiality impact but high integrity and availability impacts due to root-level command execution. This could allow an attacker to take full control of the device, disrupt its operation, or pivot into connected networks. No known exploits have been reported in the wild yet, and no patches are currently available. The vulnerability was published shortly after being reserved in late September 2025, indicating recent discovery. Unitree devices are advanced robotic platforms often used in research, industrial automation, and potentially security-sensitive environments, making this vulnerability particularly concerning due to the elevated privileges and potential for lateral movement within networks.

For European organizations deploying Unitree robots, especially in industrial, research, or automation contexts, this vulnerability poses a significant risk. Exploitation could lead to full compromise of the robotic device, enabling attackers to disrupt operations, steal sensitive data, or use the device as a foothold to infiltrate broader enterprise networks. Given the root-level access gained, attackers could manipulate device behavior, cause physical damage, or disable safety features. The impact on availability and integrity is high, potentially halting critical automated processes. Confidentiality impact is lower but still present due to possible data exfiltration. The requirement for network adjacency limits remote exploitation but does not eliminate risk in environments where these devices are connected to internal or wireless networks accessible by attackers. European organizations with connected robotic deployments must consider this a serious threat to operational continuity and security posture.

Immediate mitigation steps include isolating affected Unitree devices on segmented networks with strict access controls to limit attacker adjacency. Network monitoring should be enhanced to detect anomalous command execution or unusual Wi-Fi configuration changes. Since no patches are currently available, organizations should avoid exposing the Wi-Fi management interfaces externally and disable unnecessary wireless management features if possible. Implement strict input validation and sanitization on any custom scripts or interfaces interacting with these devices. Engage with Unitree support channels to obtain timelines for official patches or firmware updates. Additionally, consider deploying host-based intrusion detection on the devices if supported, and maintain robust network segmentation to prevent lateral movement. Regularly audit device configurations and logs for signs of exploitation attempts. Finally, prepare incident response plans specific to robotic device compromise scenarios.