WatchGuard Agent CVE-2026-6787

Local access to Windows system with WatchGuard Agent installed and standard authenticated user privileges (PR:L) are required. The attacker must be able to execute code locally, either through an existing foothold from phishing/malware or physical access scenarios. Specific configuration of WatchGuard Agent is not a limiting factor based on the CVSS vector (AT:N indicates no special attack requirements), meaning default installations are vulnerable. No user interaction is required (UI:N) once the attacker has local access. The vulnerability cannot be exploited remotely over the network without first establishing local access through other means. Limiting factors include: Windows platform requirement (not cross-platform), need for local authenticated session (remote exploitation not feasible without initial compromise), and scope constraint (SC:N/SI:N/SA:N) meaning the impact remains within the agent process rather than enabling full system compromise directly. Organizations without WatchGuard Agent deployed or those running exclusively non-Windows endpoints are not affected.

Real-world risk is moderate-to-high for organizations using affected WatchGuard Agent versions, though actual exploitation probability appears low based on available signals. The CVSS v4.0 score of 8.5 accurately reflects the local attack vector (AV:L) requiring low privileges (PR:L) and low complexity (AC:L), meaning any local user with standard credentials can potentially exploit this without specialized skills. However, no EPSS score is available in the provided data, and the vulnerability is not listed in CISA KEV, indicating no confirmed active exploitation campaigns at this time. The absence of public POC code further reduces immediate weaponization risk. From an SSVC perspective: exploitation status is none (no public POC or active exploitation reported), automatability is likely yes (hard-coded keys enable repeatable exploitation), and technical impact is high (process injection with confidentiality/integrity compromise). The primary risk is insider threats, compromised low-privilege accounts, or malware already present on endpoints leveraging this to disable or manipulate security agent protections. Organizations running WatchGuard Agent in high-security environments (financial services, healthcare, defense) should prioritize remediation due to the agent's privileged security role, even without evidence of widespread exploitation.

A local authenticated attacker with standard user credentials on a Windows workstation running vulnerable WatchGuard Agent extracts the hard-coded cryptographic key from the agent binary through static analysis or memory dumping. Using this key, the attacker crafts malicious code signed or encrypted with the same key, then injects it into the running WatchGuard Agent process, bypassing the agent's self-protection mechanisms. The injected code executes with the agent's security context, enabling the attacker to disable endpoint protections, exfiltrate credentials or security policies cached by the agent, or manipulate VPN authentication decisions. No public POC is currently available, but the low attack complexity (AC:L) suggests exploitation does not require advanced technical skills once the hard-coded key is identified.