8.6
CVSS 3.1 base score Cisco assigned to CVE-2026-76418
0
workarounds Cisco lists that address the flaw
78 days
since CISA listed the last exploited ASA VPN crash
14
ASA and FTD flaws already in CISA's exploited catalog
When the stopgap is weaker login, upgrade
Cisco published an advisory for CVE-2026-76418 on 28 Oct 2026, and the most important sentence in it is not about the bug. It is about the mitigation. To stop an unauthenticated attacker from crashing a Secure Firewall ASA or FTD through its remote access SSL VPN, Cisco's only stopgap is to switch off secondary authentication on the tunnel group. Cisco's own words on what that costs: "the security posture is lowered for remote access VPN."
That is the trade every team running these firewalls is being offered this week. Keep the second login step and leave a VPN that anyone on the internet can knock over, or remove the second step and keep the VPN up with a weaker front door. Cisco says its PSIRT became aware of active exploitation in October, so the first option is not theoretical. Neither option is a good place to stay.
Our argument is short. When the stopgap for an availability bug is weaker authentication, the stopgap is not a plan. The upgrade is the plan, and it belongs on this week's change calendar, not the next quarterly firmware cycle. The mitigation is for the narrow case where a device really cannot be upgraded in time, and even then it should be scoped, logged and time limited, with someone owning the date it comes back out.
There is a second argument underneath the first. This is the second exploited, pre authentication crash in the ASA SSL VPN in 78 days. CISA listed the previous one, CVE-2026-20349, on 11 Aug 2026. A firewall that terminates the remote workforce is a single point of failure for the business, and attackers have now shown twice since the summer that they will spend effort on making it fall over. If your remote access design has one VPN head end and no plan for it being down, the patch fixes this bug and leaves that design problem exactly where it was.
What is CVE-2026-76418 in Cisco ASA and FTD?
The Cisco advisory, first published at 19:15 GMT on 28 Oct, covers Cisco Secure Firewall Adaptive Security Appliance (ASA) Software and Secure Firewall Threat Defense (FTD) Software. It says the flaw "could allow an unauthenticated, remote attacker to cause the device to reload unexpectedly, resulting in a denial of service (DoS) condition."
The root cause, in Cisco's words, is "logic errors when secondary authentication is configured on a tunnel group." Secondary authentication, sometimes called double authentication, makes a remote user pass a second authentication server group after the first. Many organizations put their second factor there. The bug lives in the code that handles that configuration on the VPN login path, and the FTD section of the advisory removes any doubt about who can reach it: "A successful VPN login is not required to reach the affected processing path."
Cisco scores it 8.6 under CVSS 3.1. The vector is network reachable, low complexity, no privileges, no user interaction, with a changed scope and high impact on availability. Confidentiality and integrity are rated none. Read plainly, that says the attacker gets no data and no code execution from this bug. What they get is the ability to restart the firewall whenever they like.
-
Product //
Cisco Secure Firewall ASA and FTD Software, remote access SSL VPN -
Condition //
SSL VPN enabled on an interface and secondary authentication on a tunnel group -
Impact //
Unauthenticated device reload, denial of service -
CVSS //
8.6 (3.1, Cisco), availability only -
Workaround //
None; mitigation disables secondary authentication -
Exploited //
Yes, per Cisco PSIRT, October 2026
Cisco also lists what is not affected: Cloud delivered Firewall, cloud delivered Firewall Management Center, Cisco Secure Access and Secure Firewall Management Center Software. The management center is not the target here. The appliances it manages are.
One detail in the advisory deserves a moment. Cisco says the vulnerability was found during the resolution of a Cisco Technical Assistance Center support case. Cisco does not say whose case or what the customer saw. But a crash bug that comes to light through a support ticket usually means somebody had a firewall restarting and asked why. We would not read more into it than that, and neither should anyone else, but it fits the exploitation statement rather than contradicting it.
Is CVE-2026-76418 being exploited?
Yes, and we hedge exactly as far as Cisco does. The advisory's exploitation section reads: "In October 2026, the Cisco PSIRT became aware of active exploitation of this vulnerability." That is the whole statement. Cisco names no actor, no campaign and no victim count. Reporting on the advisory, including Cyber Security News, notes the same absence and that there is no reported data theft or code execution, which matches the CVSS vector.
CVE-2026-76418 was not in CISA's Known Exploited Vulnerabilities catalog in the version published on 28 Oct, the same day as the advisory. Given Cisco's statement, we expect it to be added. The catalog already holds 14 entries for ASA and FTD, going back to flaws from 2014 and 2016, and three of them were added in the last thirteen months.
The most relevant of those is CVE-2026-20349. According to The Hacker News, Cisco described it as insufficient error checking when processing HTTP requests to the Remote Access SSL VPN service, rated 8.6, letting an unauthenticated remote attacker reload the device. Cisco became aware of exploitation in August. CISA added it to the catalog on 11 Aug 2026 and gave federal agencies until 14 Aug to fix it.
Put the two side by side and the pattern is hard to miss. Same products, same service, same outcome, same score, no login needed, both exploited before or around disclosure, 78 days apart. Different root causes, which matters for patching and does not matter for planning. Somebody wants the ASA SSL VPN to stop working, and they keep finding ways.
Before those two, CISA added CVE-2025-20333 and CVE-2025-20362 on 25 Sep 2025 with a one day deadline. Those were not crash bugs, and a one day deadline tells you how seriously CISA took them. They are the reason a crash on an ASA should never be treated as only an availability ticket. We wrote about another Cisco firewall product being exploited quietly for weeks in our Interlock analysis. The lesson from both is the same: the firewall is a target in its own right, not just the thing standing in front of the targets.
Why the mitigation is the wrong trade
A denial of service bug invites a shrug. Nothing is stolen, the box comes back, the change board asks whether it can wait. On a remote access VPN head end, that shrug is a mistake, and the mitigation Cisco offers makes the trade explicit.
Figure 01
Four ways out of CVE-2026-76418. One of them is clean.
ASA SSL VPN
secondary authentication on
-
UPGRADE
fixed release, flaw gone
-
DISABLE SECONDARY AUTH
stays up, weaker login
-
DO NOTHING
reloads whenever attackers choose
-
WATCH FOR RELOADS
tells you, stops nothing
Start with what a reload means in practice. When an ASA or FTD restarts, every remote access session it terminates drops. Users reconnect when it comes back, and if the attacker sends the request again, they drop again. On a single appliance that is also the internet edge, a reload interrupts everything passing through it, not just the VPN. In a failover pair the standby takes over, which is the point of the pair, but the standby runs the same software with the same configuration and has the same flaw.
Now look at the mitigation. Cisco gives two forms of it. If the default tunnel group does not use secondary authentication, the mitigation is to turn off the tunnel group list, so users can no longer choose a group that has it. If the default group does use it, the mitigation is to remove the secondary authentication server group outright, and that is where Cisco adds the warning about a lowered posture.
On the engagements we run, when an ASA tunnel group has secondary authentication configured, it is often where the second factor lives: a first server for the directory password and a second for a one time code, or a certificate plus a password. Removing it does not leave the VPN open. It does leave the VPN guarded by whatever is in the first server, which is often a password alone. We have argued before, in our MFA analysis, that your security is set by the weakest path you still allow. Turning off secondary authentication, even for a week, creates exactly that weaker path on the most exposed login you have.
That is why the decision tree has only one clean branch. Doing nothing leaves the business offline on an attacker's schedule. Watching for reloads tells you when it happens but stops nothing. The mitigation trades an availability problem for an authentication problem, and the authentication problem is the one attackers know how to turn into access. The upgrade removes the flaw and keeps the login as strong as it was.
If a device genuinely cannot be upgraded this week, the mitigation is still better than nothing, provided it is treated as an exception. Use the first form, the tunnel group list change, wherever the default group allows it, since it keeps secondary authentication on the groups that have it. Write down the date it was applied and the date it comes out. Watch the VPN logins on that device more closely while it is in place, because that is when a password spray or a credential stuffing run has its best chance.
Which firewalls are affected by CVE-2026-76418?
Two conditions have to be true together, according to the advisory. The device must run a vulnerable release of ASA or FTD Software with the remote access SSL VPN enabled on at least one interface. And secondary authentication must be configured in the tunnel group scope.
For ASA, Cisco's guidance is to check the running configuration. If it contains the webvpn keyword with SSL VPN enabled on an interface, the first condition is met. Then check every tunnel group the tunnel group list exposes, including the default group, DefaultWEBVPNGroup, for a secondary authentication server group. If the tunnel group list is not enabled, check the default group. For FTD, Cisco points to the remote access VPN policy in Secure Firewall Management Center or Secure Firewall Device Manager, which is where those settings live on a managed device.
The fixed releases Cisco lists are:
- ASA Software: 9.16.4.104, 9.18.4.96, 9.20.4.50, 9.22.3.200, 9.23.1.48, 9.24.1.27 and 9.24.10.1.
- FTD Software: 7.0.11, 7.2.13, 7.4.9, 7.6.7, 7.7.14, 10.0.3 and 10.1.1.
Use Cisco's Software Checker against your exact train before you schedule the change, because the right fixed release depends on where you are now, and because the checker will also show whether the release you land on closes CVE-2026-20349 and the other ASA advisories you may have deferred. Upgrading for this flaw and staying exposed to the August one would be a poor use of a maintenance window.
Inventory is where this usually goes wrong. Most organizations know where their primary VPN head end is. Fewer know about the second pair stood up for a merger, the appliance at a regional office that still terminates a handful of contractors, or the lab firewall someone exposed to the internet to test a client profile. Each of those is an ASA or FTD answering on the internet with a VPN login, and each one is in scope for this advisory.
How we test VPN edge resilience on an engagement
On a sanctioned external test, the VPN head end is one of the first things we map, because it is the one system every remote employee depends on and every attacker can reach. The work is mostly discovery and configuration review, and almost none of it involves knocking anything over.
We start by finding every VPN login the client exposes, not just the one they told us about. Our external recon method, written up on the external recon page in our Learn section and automated in part by our open source atarus-recon tool, maps host names and addresses from DNS, certificate transparency and the ranges in the scope letter. VPN portals are easy to recognize once found. The useful finding is usually a portal the client did not know was still live.
Then we compare what each portal offers with what the client intends it to offer. Which tunnel groups can a user choose from the login page? Which of them require a second factor, and which quietly do not? Is the default group configured the way the client believes it is? Those questions matter for this advisory, and they matter more for the next credential attack, because the weakest selectable group is the one an attacker will use.
We do not crash production firewalls to prove a denial of service. For a bug like CVE-2026-76418, the release, the configuration and the reachability of the login service are enough evidence to rate the finding. Our network impact without damage guide sets out how we prove impact on a network test without causing it. Where a client wants to see resilience tested for real, we agree a lab or a maintenance window, a rollback plan and a named contact before anything is sent.
Finally, we ask the question the advisory does not: what happens to the business when the VPN is down for an hour? Is there a second head end in another location? Do critical staff have another way in that is just as well protected, or does the fallback turn out to be a remote desktop gateway with a password? An attacker who can crash the front door on demand will be very interested in whatever the business opens instead.
What is still coming for the ASA VPN
Two exploited pre authentication crash bugs in one service in 78 days is not a coincidence we would plan around. The ASA SSL VPN is a large, old, internet facing parser that sits on a very large number of networks, and attackers have shown since the summer that they are probing it for ways to make it fall over. We expect more findings in the same code, from researchers and from whoever found these two.
We also expect the motive to stay unclear for a while, and we will not guess at it in public. Cisco has not named an actor for either flaw. What we can say from the defender's side is that a reliable way to reload a perimeter firewall is useful for more than one purpose: an outage on its own, pressure during an extortion, a distraction while something else happens, or a way to push users onto a weaker access path. None of those require the attacker to explain themselves. All of them are reasons to patch quickly and to plan for the VPN being unavailable.
The structural fix is the same one we made in our edge week analysis. Edge devices keep returning to the exploited catalog, and each advisory is one more reason to treat them as tier zero: inventoried, upgraded on an emergency path rather than a quarterly one, watched like a domain controller, and backed by a second route for the people who depend on them. Patch CVE-2026-76418. Then fix the plan that made a crash on one appliance a business outage.
What to do today
-
Upgrade every ASA and FTD running the SSL VPN to a fixed release this week. Use Cisco's Software Checker against your current train and land on a release that also closes CVE-2026-20349.
-
Find every VPN head end you expose, not just the main one. Regional offices, merger leftovers and lab boxes all count, and each one answering on the internet is in scope.
-
Use the mitigation only as a dated exception. Prefer disabling the tunnel group list over removing secondary authentication, record when it goes in and when it comes out, and watch VPN logins closely while it is in place.
-
Alert on unexpected firewall reloads. Treat an unplanned restart of a VPN head end as a security event to investigate, not a hardware ticket.
-
Plan for the VPN being down. Decide now how critical staff get in if the head end is unavailable, and make sure that route is protected by the same second factor.
SOURCES
- Cisco Security Advisory // CVE-2026-76418, the 8.6 score, active exploitation, the mitigation and its warning, fixed releases
- CISA KEV catalog // the 14 ASA and FTD entries and CVE-2026-20349 added on 11 Aug 2026
- The Hacker News // CVE-2026-20349, the August SSL VPN reload flaw, and its federal deadline
- Cyber Security News // the fixed ASA and FTD releases and that no attacker or victim was named