Hackers Exploit Cisco Expressway Vulnerability After PoC Release
Attackers leveraged a critical Cisco Expressway vulnerability (CVE-2026-76466) shortly after proof-of-concept exploit code became publicly available, breaching a Russian organization.

Cybersecurity firm F6 has reported that malicious actors exploited a critical vulnerability in Cisco Expressway Series software, specifically CVE-2026-76466, shortly after proof-of-concept (PoC) exploit code was released publicly. The attackers successfully breached a Russian organization by compromising clustered Expressway devices, which are crucial for enabling secure remote collaboration services such as video, voice, and instant messaging.
The vulnerability, carrying a CVSS score of 9.8, allows unauthenticated remote attackers with network access to execute commands with root privileges. It affects both virtual and appliance versions of Cisco Expressway Core and Expressway Edge, but only when deployed in a cluster. Cisco's advisory indicates the flaw resides in the Expressway database service feature and is related to persistent cookies, enabling attackers to issue RPC commands and gain root-level control over the underlying operating system.
Cisco released a patch for this vulnerability in Expressway X15.5.20, stating that there are no workarounds and that customers on earlier versions must upgrade. Despite Cisco's initial advisory noting no awareness of public announcements or malicious use of the flaw, a PoC exploit subsequently appeared on GitHub, followed closely by the attack investigated by F6.
The attackers targeted two internet-exposed Expressway X15.5.3 devices and then pivoted to two internal cluster nodes within the victim's network. After gaining root privileges, they deployed custom Python scripts designed for lateral movement. These tools were intended to scan the network, locate SMB resources, interact with SQL Server and WinRM, and search GitLab repositories for credentials and sensitive information.
F6 observed that the attackers dynamically modified the scripts during the attack, noting that the source code contained comments and fix descriptions suggestive of AI assistant usage. The firm did not identify the victim or attribute the attack to a specific threat actor.
To detect exploitation, F6 recommends examining the Erlang system log on Expressway devices. The firm has also published associated IP addresses and file paths in a Russian-language press release. While Cisco offers no workarounds, F6 suggests that organizations unable to patch immediately should block external access to TCP ports 4371–4372 at the firewall, permitting connections only from trusted cluster nodes.
This incident highlights the rapid exploitation of vulnerabilities once PoC code becomes available, underscoring the critical need for organizations to apply patches promptly, especially for internet-facing systems. The use of AI-assisted tools in crafting attack scripts also points to an evolving threat landscape where attackers are leveraging new technologies to enhance their capabilities.
This new report details the specific post-exploitation activities observed, including the use of custom Python scripts for network scanning, lateral movement, and interaction with SQL Server and WinRM. Analysis of these scripts suggests potential development with AI assistance, indicated by embedded comments and fix descriptions. The attackers successfully breached four devices, gaining access to the internal network and exfiltrating credentials and sensitive data.