Parallels RAS Client CVE-2026-18262

The attacker must already possess low-privileged local code execution on the target Windows machine - this is an explicit prerequisite per the CVSS AV:L/PR:L vector and the vulnerability description. The Parallels RAS Client must be installed and the RAS RDP Backend Service must be actively running, which is part of the standard product installation. No network-based or remote exploitation is possible per AV:L - access to the local system is mandatory. No special non-default configuration, specific protocol, or elevated credential is needed beyond the local low-privileged foothold. Exploitation is limited to systems where Parallels RAS Client 21.0.26296 is deployed; systems without this specific product are unaffected.

The CVSS 3.0 vector AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H accurately describes a reliable, low-complexity local privilege escalation path: the only prerequisite is an existing low-privileged foothold on the target host, with no interaction barriers or elevated configuration requirements. No EPSS data was provided, which limits probabilistic forecasting, but the straightforward CWE-749 class and absence of complexity barriers suggest exploit development is not technically demanding once the service interface is mapped. No CISA KEV listing and no confirmed public POC have been identified at time of analysis, meaning active widespread exploitation is not currently documented - however, even a mid-sized municipal IT team in ustka running Parallels RAS environments should treat this as urgent, since any attacker with a local foothold (delivered via phishing, malware, or insider threat) could immediately escalate to full SYSTEM control. The unchanged scope (S:U) confirms impact is contained to the compromised host rather than adjacent systems, which slightly limits blast radius in segmented environments but does not diminish local severity.

An attacker with a low-privileged local account on a Windows system running Parallels RAS Client 21.0.26296 - obtained, for example, through phishing or a separate vulnerability - locates and invokes the exposed function in the RAS RDP Backend Service using crafted parameters or a direct service call. Because the service operates as SYSTEM and performs no access control check on the caller, the invoked function executes arbitrary code in the SYSTEM context. No public exploit code has been confirmed, but CWE-749 class vulnerabilities are typically straightforward to exploit once the service interface is reverse-engineered.