Configuration change as a mitigation for a vulnerability

A configuration change removes a precondition the flaw needs, while the code stays. Cisco's advisory for CVE-2026-76418 (28 Oct 2026) lists no workaround but one mitigation: no tunnel-group-list enable, or no secondary-authentication-server-group when the default group uses secondary authentication. Read 19 Nov 2026.

What this control does toward each outcome

  1. RemediateLeaves the flaw: The underlying flaw stays. Microsoft defines a workaround as a change that 'does not correct the underlying vulnerability'.
  2. Mitigate now, fix laterLowers exposure: Changes a setting so the vulnerable behavior is not reachable or not exploitable under default use.Guidance points here
  3. Accept with a dated exceptionIf the fix date slips: accept the residual risk in a signed, dated, reviewed exception.Also allowed

Decision Table row: A control already sits in front of the vulnerable component (WAF rule, segmentation, access restriction). Group: The code path or setting.

Evidence to keep: Approved configuration baseline entry; Read-back output with date; Drift alert rule; Approved deviation record if the baseline is changed.

A setting counts when it removes a precondition

Vendors list the conditions a flaw needs. A configuration change is a mitigation when it makes one of those conditions false and keeps it false. It is not one when it only narrows the condition. Vendors also use the words differently. Cisco's advisory for CVE-2026-76418 says "There are no workarounds that address this vulnerability. However, there is a mitigation". Microsoft's bulletin MS10-046 (2010) used "workaround" for a change you make and "mitigation" for a setting "existing in a default state", so the labels do not carry across vendors. The vocabulary is settled on workaround vs mitigation.

Settings and conditions from the CVE records and vendor advisories read 19 Nov 2026.
CVESettingNamed byDoes it remove the precondition?
CVE-2026-76418, Cisco ASA and FTD (KEV 29 Oct 2026)no tunnel-group-list enable, or no secondary-authentication-server-groupCiscoYes. The flaw needs secondary authentication on a tunnel group in scope.
CVE-2026-123158, OpenSSH on musl without PAM (KEV 9 Nov 2026)PasswordAuthentication no, including inside Match blocksOpenSSH and AlpineYes. The flaw needs password authentication on. The cost is that nobody can log in with a password.
CVE-2026-109618, PHP out-of-memory corruption (KEV 26 Oct 2026)memory_limit=-1, then restart every workerPHPOnly in part. PHP says a failed mmap() can still reach the vulnerable path, and a runaway script is no longer stopped.

The named case: CVE-2026-76418 on Cisco ASA and FTD

Cisco published the advisory at 19:15 GMT on 28 Oct 2026. A logic error in the Remote Access SSL VPN service lets an unauthenticated attacker reload the device, a denial of service rated 8.6 on CVSS 3.1. The device is affected when it runs a vulnerable release, the SSL VPN login service is reachable, and secondary authentication is configured in tunnel group scope. Cisco's advisory dates exploitation only as "October 2026" and gives no day, so the first dated public points are the advisory on 28 Oct and the CISA KEV entry on 29 Oct 2026, due 1 Nov 2026.

Cisco's mitigation has two cases. Case 1: if DefaultWEBVPNGroup does not use secondary authentication, "enter no tunnel-group-list enable" in WebVPN configuration mode. Users then cannot pick a connection profile from a list and "will need to know the correct profile to use". Case 2: if DefaultWEBVPNGroup uses secondary authentication, use no secondary-authentication-server-group to remove it. Cisco states the cost: "the security posture is lowered for remote access VPN". It adds that the mitigation was proven in a test environment and that customers "should not deploy any workarounds or mitigations before first evaluating the applicability to their own environment".

Cisco advisory cisco-sa-asaftd-tunnel-dos-s6mgkb75 (28 Oct 2026), read 19 Nov 2026.
ProductTrainFirst fixed release
ASA9.16 and earlier9.16.4.104
ASA9.189.18.4.96
ASA9.209.20.4.50
ASA9.229.22.3.200
ASA9.239.23.1.48
ASA9.249.24.1.27 (9.24.10.1 also listed)
FTD7.0 and earlier7.0.11
FTD7.27.2.13
FTD7.47.4.9
FTD7.67.6.7
FTD7.77.7.14
FTD10.010.0.3
FTD10.110.1.1

A worked read-back that catches the wrong fix

The device below is illustrative, running ASA 9.20.4.40, below the first fixed release 9.20.4.50. The facts about Cisco's note are real: Cisco says that "disabling the group list does not remove exposure when that default group uses secondary authentication".

  1. Thu 29 Oct 2026: KEV entry, due Sun 1 Nov 2026. Read show running-config webvpn and confirm SSL VPN is enabled on an interface.
  2. Read show running-config tunnel-group. Search every group, including DefaultWEBVPNGroup, for secondary-authentication-server-group. Suppose 2 of 5 groups have it, one of them the default group.
  3. An engineer sees tunnel-group-list enable and runs no tunnel-group-list enable. The login page no longer lists profiles.
  4. Read show running-config tunnel-group again. The default group still has secondary-authentication-server-group, so the device is still exposed. The command changed the login page, not the precondition.
  5. Apply Case 2: remove secondary authentication from the default group. Remove it from the other group, since the aim is none in scope. Read both groups back.
  6. Record the lowered posture, because VPN users are no longer asked for the second credential, and set the end date to the upgrade to 9.20.4.50.

The baseline rule that governs it

SP 800-53 CM-6 asks organizations to "Establish and document configuration settings for components employed within the system that reflect the most restrictive mode consistent with operational requirements" (item a) and to "Identify, document, and approve any deviations from established configuration settings" (item c). Removing secondary authentication is a deviation from the standard: it lowers the posture. It needs an approver and a date, or it is a private change an assessor cannot accept. Item d asks that changes to settings are monitored and controlled, which is the drift rule below.

Check that it took, and what it costs

  1. Read the setting back from the running configuration and compare it with the approved value, as in the worked case.
  2. Run the vendor's own check for the condition if one exists (here the configuration checks in Cisco's advisory).
  3. Rebuild a test instance from the standard image and confirm the setting is present, since a rebuild from an old image brings the default back.
  4. Alert on drift from the approved value so a revert opens a ticket.

The cost is what the setting switches off. Cisco's cases cost either the profile list on the login page or the second credential. The OpenSSH setting costs all password logins; test key login in a second session first. Revisit the entry when the fix is installed and the setting can go back, or when exploitation is reported against devices that already carry the setting, in which case it was never enough.

Record to keep

Configuration change record (deviation from baseline)

  • Finding: CVE-2026-76418, Cisco ASA and FTD (Cisco advisory 28 Oct 2026)
  • Asset and release: [hostname], [release]; first fixed release [release]
  • Precondition removed: secondary authentication in tunnel group scope
  • Setting changed: [command], on [groups]; before [value]; after [value]
  • Read-back on [date]: [output summary]; default group checked: [yes or no]
  • Posture cost accepted: [what was lowered]; approver [name]
  • Baseline: added to baseline [yes or no]; deviation ID [ID]
  • Drift alert rule: [rule ID]
  • Rebuild test: [date], setting present: [yes or no]
  • KEV due date: [date]; met: [yes or no]
  • End date: [date of upgrade]; revert step: [command]; owner [name]

Sources

Last reviewed