CVE-2026-76466 - Cisco Expressway Series Use of Hard-coded Credentials Vulnerability

CVE-2026-76466

Cisco Expressway - Unauthenticated Root RCE via Hard-coded Cluster Database Credentials

What is Cisco Expressway?

Cisco Expressway, descended from the TelePresence Video Communication Server, is the edge gateway for Cisco collaboration deployments. It comes in two roles: Expressway-E sits in the DMZ or directly on the internet, and Expressway-C sits inside the network and traverses the firewall to reach it. Together they carry SIP and H.323 video calls, Mobile and Remote Access for Jabber and Webex clients, and business-to-business calling, without requiring a VPN.

That design puts Expressway-E on the public internet by definition; it is the box an organisation deliberately exposes so remote users and partner organisations can call in. It is also trusted inside: it holds collaboration credentials, terminates sessions for internal users, and sits on a network segment with reach into the voice and directory infrastructure. For high availability, Expressway nodes are deployed in clusters of up to six, and the cluster members replicate configuration through a database service. That cluster database is where this vulnerability lives.

Overview

CVE-2026-76466 is a hard-coded credentials flaw, CWE-798, in the Expressway database service. An unauthenticated remote attacker who can reach the cluster database port, and who holds the right cookie, can connect and invoke RPC commands that execute arbitrary operating system commands as root. Cisco rates it 9.8 critical with the vector AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H, tracked internally as bug CSCww27732.

Cisco published advisory cisco-sa-vcs-static-cred-hr7EWZ8t on 2026-10-21 at 16:00 GMT, alongside a broader October 2026 Expressway hardening release. At publication, Cisco PSIRT stated it was not aware of public announcements or malicious use. That changed quickly. A public proof-of-concept exploit appeared on GitHub on 2026-11-05, and on 2026-11-19 the security firm F6 reported the first investigated case of exploitation in the wild. CISA added the CVE to the KEV catalog the following day with a 2026-11-23 deadline.

Cisco's advisory limits the vulnerability to devices "configured in a cluster." A standalone Expressway is not affected by this CVE, though Cisco still directs standalone operators to X15.5.20 for the other issues in the same hardening release.

Affected Versions

Product Vulnerable Fixed
Expressway-C (clustered, virtual or appliance) X15.5.3 and earlier X15.5.20
Expressway-E (clustered, virtual or appliance) X15.5.3 and earlier X15.5.20
Expressway / TelePresence VCS X14 and earlier No fixed X14 build; migrate to a fixed release
Expressway (standalone, not clustered) Not affected by this CVE X15.5.20 recommended for the October 2026 hardening CVEs

Technical Details

Hard-coded credentials means a secret baked into the shipped software rather than generated per installation, so the same value works on every deployment worldwide. Cisco's own description of the cause is narrow: the use of persistent cookies, combined with insufficient client-certificate validation when the TLS Verify Mode for clustering is set to Enabled. An attacker in possession of those cookies connects to the database service and issues specific RPC commands.

Public analysis by Ostorlab goes further than Cisco does, and these details should be read as third-party research rather than vendor-confirmed. According to that analysis, the Expressway cluster database runs on Erlang/OTP with a distribution listener on TCP 4372 (4371 is also a cluster port). Erlang distribution authenticates peers with a shared secret called a cookie. The cm_server module adds a maintenance node identity, [email protected], to the set of allowed nodes and installs a matching cookie, both compiled into the image and therefore identical everywhere. An attacker opens a TLS connection to the distribution port, completes the Erlang handshake as that maintenance node using the static cookie, and then issues rpc:call to os:cmd. That is the whole chain: one connection, no credentials the attacker did not already know, and the Erlang runtime executes shell commands as root because that is the user the service runs as.

One inconsistency is worth noting. Cisco ties the client-certificate weakness to TLS Verify Mode set to Enabled, while Ostorlab reports testing the proof of concept against a two-node X15.3.2 cluster running in Permissive mode. Do not assume a particular TLS Verify Mode setting protects you.

Discovery

Cisco credits Stefan Aleksic of Jane Street with reporting the vulnerability. The public proof of concept was published independently roughly two weeks after the advisory, by the GitHub user dinosn, in a repository created on 2026-11-05; one secondary source instead dates public PoC availability to 2026-11-12, but two independent write-ups agree on the earlier repository date. Ostorlab published the detailed Erlang-level analysis, and F6 published the first exploitation report.

Exploitation Context

F6 reported on 2026-11-19 that it had investigated an intrusion in which attackers compromised two internet-facing Expressway devices and then pivoted to two further cluster nodes on the internal network, four devices in total, at an organisation in Russia. Secure.com reports all four were running X15.5.3. After gaining root, the attackers ran custom Python scripts that scanned for SMB shares, probed SQL Server and WinRM, and searched GitLab repositories for credentials: the behaviour of an actor using the gateway as a beachhead for internal reconnaissance rather than targeting the video infrastructure itself. F6 did not attribute the activity to a named group.

The sequence is the notable part. The gap from patch to first reported breach was about four weeks, and only two weeks from the public proof of concept. Cisco's advisory still carried its "not aware of malicious use" language as of 2026-11-20, which is a reminder that a vendor's exploitation status reflects what the vendor knows, not the current state of the internet. No published count of internet-exposed Expressway devices exists for this CVE, though Expressway-E is internet-facing by design, so the exposed population is substantial. The F6 incident is currently a single-source report, but the combination of a public exploit, a static credential and an internet-facing appliance makes further activity near-certain.

Remediation

  1. Upgrade clustered Expressway-C and Expressway-E to X15.5.20 or later. The BOD 26-04 deadline is 2026-11-23. There is no workaround: Cisco states explicitly that none addresses this vulnerability.
  2. If you run X14 or earlier, there is no fix on that train. Migrate to X15.5.20 or later.
  3. Patch standalone Expressway nodes to X15.5.20 as well. They are not affected by this CVE, but they are in scope for the other issues in the October 2026 hardening release.
  4. Restrict the cluster ports, TCP 4371 and 4372, to cluster peers only, at the host firewall and the network. This reduces exposure and is good practice regardless, but it is not a substitute for patching.
  5. Hunt for compromise on any device that was unpatched and reachable. Check System > Status for cluster peers you do not recognise. Search the log bundle, including erlang.log, for the string [email protected]. Look for Python files under /mnt/harddisk/ whose names match a leading "i" followed by four digits and an optional letter. Review firewall logs, including /var/log/firewall.log, for connections to the reported addresses 66.179.31.32 and 18.130.47.116. Treat any hit as a lead to investigate, not proof on its own.
  6. If you find evidence of access, assume root on the appliance and a staging point for internal reconnaissance. Rotate collaboration and directory credentials the device could reach, review SMB, SQL Server, WinRM and GitLab authentication logs from the affected period, and rebuild the appliance from a known-good image rather than patching it in place.

Key Details

PropertyValue
CVE ID CVE-2026-76466
Vendor / Product Cisco - Expressway Series
NVD Published2026-10-21
NVD Last Modified2026-11-20
CVSS 3.1 Score9.8
CVSS 3.1 VectorCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
SeverityCRITICAL
CWE CWE-798 find similar ↗
CISA KEV Added2026-11-20
CISA KEV Deadline2026-11-23
Known Ransomware Use No

CVSS 3.1 Breakdown

Attack Vector
Network
Attack Complexity
Low
Privileges Required
None
User Interaction
None
Scope
Unchanged
Confidentiality
High
Integrity
High
Availability
High

Required Action

CISA BOD 26-04 Deadline: 2026-11-23. Apply mitigations in accordance with vendor instructions, ensuring compliance with CISA's BOD 26-04 Prioritizing Security Updates Based on Risk (see URL in Notes) guidance and CISA's "Forensics Triage Requirements" (see URL in Notes). Follow applicable BOD 26-04 guidance for cloud services or discontinue use of the product if mitigations are unavailable. Stakeholders are responsible for evaluating each asset's internet exposure and ensuring adherence to BOD 26-04 patching guidelines.

Timeline

DateEvent
2026-10-21Cisco publishes advisory cisco-sa-vcs-static-cred-hr7EWZ8t and the X15.5.20 fixed release; CVE record published
2026-11-05Public proof-of-concept exploit published to GitHub
2026-11-19Security firm F6 reports the first observed in-the-wild exploitation
2026-11-20Added to CISA Known Exploited Vulnerabilities catalog
2026-11-23CISA BOD 26-04 remediation deadline