How to Check and Patch Cisco SD-WAN Manager CVE-2026-76504
Cisco Catalyst SD-WAN Manager CVE-2026-76504 is an actively exploited critical authentication bypass. Learn how to check logs, preserve evidence, restrict access and patch affected Managers.
On this page
Cisco disclosed CVE-2026-76504 on September 30 after its security team became aware of active exploitation against Catalyst SD-WAN Manager. The critical authentication bypass can let an unauthenticated remote attacker reach the Manager with administrator privileges, and Cisco says there is no workaround that fixes the vulnerability. If you operate an affected on-premises Manager, the job is not just to install a fixed release: you should first preserve evidence and check the appliance for the exploitation patterns Cisco has published.
Check the Catalyst SD-WAN Manager release before changing anything
CVE-2026-76504 affects Cisco Catalyst SD-WAN Manager because of improper handling of URI encoding in API session authentication. An attacker can send a specially crafted HTTP request in which a character in the login path is encoded, allowing the request to bypass an authentication rule that is supposed to restrict access to the API endpoint. Cisco rates the vulnerability 9.8 on the Common Vulnerability Scoring System, with no authentication or user interaction required.
Before upgrading, record the exact software release running on each Manager. Cisco's fixed-release table covers several active release trains, so do not apply a generic upgrade instruction based only on the product name. Deployments earlier than 20.9 must migrate to a fixed release, while the supported trains have specific first-fixed versions.
| Current release train | First fixed release |
|---|---|
| Earlier than 20.9 | Migrate to a fixed release |
| 20.9 | 20.9.10.1 |
| 20.12 | 20.12.8.2 |
| 20.15 | 20.15.6.1 |
| 20.18 | 20.18.4.1 |
| 26.1 | 26.1.2.1 |
| 26.2 | 26.2.1 |
The fixed release matters because Cisco says the vulnerability affects deployments regardless of system configuration. In other words, you should not assume that a particular SD-WAN feature being unused makes an affected Manager safe. Cisco also says Cisco SD-WAN Cloud is fixed separately in release 20.15.605, with no customer action required for the Cisco-managed service.
Restrict internet access before the upgrade
Cisco does not provide a workaround for CVE-2026-76504, but it recommends restricting access to on-premises Managers from unsecured networks such as the public internet. This is a containment measure, not a substitute for installing the fixed software. A management system that can be reached only from an approved administrative network has a much smaller exposure window while the upgrade and investigation are underway.
Review the firewall or access-control rules in front of the Manager and identify every network that can reach its management services. Cisco's SD-WAN hardening guidance recommends that administrative interfaces should not be exposed directly to the internet. If remote administration is required, use the organization's established protected management path rather than opening the Manager broadly to external addresses.
Preserve evidence before restarting or upgrading
Because exploitation was already occurring before the advisory was published, check for compromise before treating the upgrade as the end of the incident. Cisco specifically recommends using the request admin-tech command on the Catalyst SD-WAN Manager before opening a Technical Assistance Center case. The resulting admin-tech file gives Cisco material to examine when determining whether the Manager was compromised.
If your organization has centralized logging, preserve the relevant records there as well. Local logs can rotate, and an upgrade or restart can change the evidence available on the appliance. Keep the original timestamps, source addresses, and surrounding authentication events so a suspicious request can be correlated with what happened immediately before and after it.
Search serviceproxy-access.log for encoded login paths
Cisco identifies /var/log/nms/containers/service-proxy/serviceproxy-access.log as one of the primary places to investigate. Look for requests involving j_security_check that came from unknown or unauthorized addresses, especially requests where a character in the path has been percent-encoded. Cisco's published example contains POST /%6a_security_check HTTP/1.1, where %6a represents the letter j.
grep -Ei '(%[0-9a-f]{2}).*security_check|security_check.*(%[0-9a-f]{2})' /var/log/nms/containers/service-proxy/serviceproxy-access.logThe exact encoded value is not the important part. Cisco says the example is only one representation and that an attacker can encode a character in the request path. Searching exclusively for %6a_security_check could therefore miss other encoded variants.
A matching request is an investigation lead, not automatic proof of compromise. Cisco warns that related entries can also occur during normal operation, so compare the source address with your approved management networks and determine whether the request occurred during legitimate administrative activity.
Check vmanage-server.log for reserved service-account activity
The second file to inspect is /var/log/nms/vmanage-server.log. Cisco specifically calls out entries associated with j_security_check and usernames beginning with viptela-reserved-. Those accounts are internal service accounts, so the useful signal is the combination of the login-path activity, the account name, the source information, and whether the event fits normal operation.
grep -F 'viptela-reserved-' /var/log/nms/vmanage-server.logNext, correlate any matching timestamp with the access log. An encoded login-path request from an unfamiliar source followed by activity involving a reserved account deserves investigation, particularly when it falls outside an approved maintenance window. Do not declare an appliance compromised from the account prefix alone; Cisco explicitly warns that some of these indicators can occur during standard operations.
Use the two logs together instead of treating one match as a verdict
The most useful investigation combines the network-facing request with the application event. For example, a request containing an encoded character in the j_security_check path can establish that someone attempted the attack path, while a related application-log entry can show which internal account was involved. The source address then provides another correlation point against firewall records, VPN logs, jump-host activity, or known administrative systems.
This matters because a successful HTTP response does not automatically demonstrate that an attacker obtained persistent control. Conversely, the absence of one exact request does not prove that a system was never compromised, especially if logs have already rotated or were not retained for the relevant period. Cisco's published indicators are therefore best used as focused hunting leads rather than as a complete forensic detection system.
Upgrade to the first fixed release for your train
Once evidence has been preserved and the initial compromise check is complete, upgrade to the first fixed release for the installed train. Cisco lists 20.9.10.1 for the 20.9 train, 20.12.8.2 for 20.12, 20.15.6.1 for 20.15, 20.18.4.1 for 20.18, 26.1.2.1 for 26.1, and 26.2.1 for 26.2. Releases earlier than 20.9 require migration to a fixed release rather than a simple patch within the same train.
Do not choose a version solely because it is numerically newer. Cisco's compatibility documentation and upgrade matrix should be used to determine the supported path for the particular SD-WAN deployment. The goal is to reach a release that Cisco identifies as fixed while preserving compatibility across the control components managed as part of the SD-WAN environment.
Verify the Manager after the upgrade
After the upgrade, confirm that the Manager reports the intended fixed release and that the SD-WAN control plane returns to its expected operating state. Check that administrators can authenticate normally, that device connectivity is healthy, and that policy and configuration management continue to work. Then repeat the log review rather than assuming the new version has erased evidence from the previous vulnerable state.
If suspicious indicators were found before the upgrade, do not close the incident merely because the software now reports a fixed version. Cisco recommends opening a Severity 3 Technical Assistance Center case with CVE-2026-76504 in the title when help is needed to determine whether the Manager was compromised, and it asks customers to provide the admin-tech file generated before remediation.
Check every Manager, not just the one that triggered the alert
Catalyst SD-WAN Manager provides centralized control over the SD-WAN environment, so an organization may operate more than one Manager or control component. Build an inventory of the affected release trains and repeat the version check, network-exposure review, evidence collection, and log hunting for each relevant appliance. Cisco's advisory and remediation guidance are written for the affected Manager software rather than a single appliance instance.
CVE-2026-76504 is now a practical incident-response problem as well as a patching problem: Cisco confirmed active exploitation in September, assigned a 9.8 severity score, and added fixed releases across the current software trains. The safest workflow is therefore straightforward: restrict exposure, preserve evidence, search the two Cisco-named logs, upgrade to the correct fixed release, and escalate suspicious findings before declaring the Manager clean.
Share this post



