CVE-2026-55978

This CVE reveals a critical design flaw in CatchPulse: the kernel filter communication port accepts connections from any local process, effectively allowing untrusted code to bypass the security policy it's supposed to enforce. If you have any local code execution ability—whether from malware, a compromised credential, or another vulnerability—you can send messages directly to the enforcement layer and disable or manipulate protection policies. The CVSS 8.4 score significantly understates the real risk in enterprise environments where CatchPulse is the primary endpoint controls. A compromised workstation in a fleet where CatchPulse enforces application whitelisting doesn't just evade one policy—it potentially seizes the enforcement layer itself, turning the security product into a weapon that whitelists malware as 'signed known good' code. Verify immediately whether the kernel filter port (check process enumeration for CatchPulse-related listening ports) permits connections from non-admin processes—this is your blast radius indicator. If you're running CatchPulse in containerized or multi-tenant environments, assess whether container workloads can reach this port; if so, the exploit path collapses from 'malware on my workstation' to 'any workload on this host,' dramatically expanding exposure. Prioritize patching aggressively—the EPSS low-probability framing is misleading because the actual exposure window isn't days but potentially years across your install base. Beyond patching, examine whether CatchPulse has compensating kernel-level authentication downstream; if the port is the only chokepoint with no secondary validation, your risk profile is worse than the CVSS suggests. This vulnerability class has recurred across endpoint security products for decades because architects treat localhost as a non-adversarial trust domain—an assumption that fails in any multi-user, BYOD, or shared-infrastructure environment.

Reviewed through automated stages and approved by a human before publication.