CISA added four records to the known exploited catalog on 9 November, all four due 12 November. The release that carries them is catalog version 2026.11.10, published on 10 November 1. Two of them are new here: in Aspera Faspex, an existing account becomes an administrator, and on OpenSSH builds that use musl without PAM, one long password gets through. The third is GPUStack before 2.2.0, which already has its own post here from 10 November. The fourth is CVE-2026-116779 in SignalWire FreeSWITCH, which I wrote about on 7 November when it held no catalog entry. It has one now, added on 9 November with the same due date. NVD recorded the three below between 4 and 6 November.

The OpenSSH bypass that needs one long password

CVE-2026-123158 lives in sshd on Linux systems running musl libc where sshd does not use PAM, before OpenSSH 10.6p2. For some account layouts, any password between 256 and 1024 characters authenticates 2. Alpine’s advisory puts the attempt at one connection, working every time, with no retries 3. The CNA scores it 8.1 on a vector with high attack complexity 2, and that reads low for a login that works on the first try.

The mechanism is a collision in a sentinel value. Without PAM, sshd hashes the attempt with crypt() and compares the result to the stored field, and when that field is too short to hold a salt, sshd borrows the salt from another account. musl’s crypt() returns the string * once the key runs past 256 bytes, and * is what the password field holds for an account with no password set. The borrowed salt plus the * result compares equal, and the login succeeds 3.

Four conditions have to line up: a sshd that does not use PAM (the Alpine openssh-server package, which is built without it, or openssh-server-pam with UsePAM no), PasswordAuthentication yes, one account whose password field is *, and one other account holding a $5$ or $6$ hash 3. That is less rare than it sounds. Root ships as root:* in the Alpine base image, key-only users created with adduser look the same, and the official tiny-cloud images from 3.18.11 through 3.24.3 carry an alpine user with a * field and passwordless doas. Those images are safe on first boot and become reachable the moment a $6$ hash exists on the machine, which a console password creates 3.

The fix is 10.6p2, and Alpine backported it to every release from 3.16 through 3.24, including 3.21 after its 1 November end of support. The advisory’s announcement names the four point releases that carry the fixed package, 3.21.10, 3.22.8, 3.23.8 and 3.24.4 3, and its table gives one for each older branch: openssh 9.7_p1-r6 on 3.20, 9.6_p1-r3 on 3.19, 9.3_p2-r4 on 3.18, 9.1_p1-r7 on 3.17 and 9.0_p1-r6 on 3.16 3. Upgrading does not restart a running sshd, so the daemon needs a restart afterwards, and PasswordAuthentication no closes the hole without the update. The advisory’s own attack check is a grep for Accepted password for against an account with no password in /var/log/messages 3.

GPUStack, already covered

CVE-2026-102895 is a 9.8 in GPUStack before 2.2.0 4, and the 10 November post covers the two ways a remote caller could make its request look local and what the administrator seat reaches once it lands. Nothing there changes with the catalog entry, so I am not retelling it here.

Aspera Faspex, from an account to an administrator

CVE-2026-123218 sits in Aspera Faspex 5.0.0 through 5.0.19, where a message crafted by an authorized user bypasses the authorization check and comes back with full administrative privileges 5. IBM scores it 9.8 on a vector that reads PR:N, while the record’s own description requires an authenticated account 5. I go by the description, so I treat the starting point as an account somebody already granted rather than an anonymous caller. 5.0.19 PL1 carries the fix 6.

Faspex is a file transfer server, so it tends to sit where partners and vendors can reach it, and the accounts on it are the ones worth upgrading. Where no external party holds an account, the exposure narrows to a phished or reused credential. That part is my reading: the catalog describes a flaw an authenticated user can use to escalate and says nothing about how that account was obtained 1.

The relay before the 12 November date

Those three carry the same 12 November due date in catalog 2026.11.10, and so does the FreeSWITCH record 1. I use that date as the point where I want an answer in my own environment, and the answer is either a fixed build or a recorded reason the product is not here.

The SSH question has a command answer, so I would take it first: hosts running a musl-based image with sshd, sshd -T for the UsePAM and password settings, and a look for accounts with * in the hash field. GPUStack and Faspex are inventory questions, and the version decides them where either one is installed. The fixed builds are GPUStack 2.2.0 and Faspex 5.0.19 PL1 46.

The findings go to the teams that own those systems as priorities, with the restart requirement attached. An upgraded package in a Thursday change window is not a closed item while the daemon still runs the old build, so the restart belongs in the same ticket as the upgrade.

None of this is urgent until I know which hosts run these products, so the host list comes first.


  1. Cybersecurity and Infrastructure Security Agency. (2026, November 10). Known exploited vulnerabilities catalog (version 2026.11.10). CISA. https://www.cisa.gov/known-exploited-vulnerabilities-catalog ↩︎ ↩︎ ↩︎

  2. National Institute of Standards and Technology. (2026, November 5). CVE-2026-123158 detail. NVD. https://nvd.nist.gov/vuln/detail/CVE-2026-123158 ↩︎ ↩︎

  3. Alpine Linux. (2026, November 5). Security advisory: openssh-server password authentication bypass. Alpine Linux. https://alpinelinux.org/posts/2026-11-05-openssh-password-authentication-bypass/ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  4. National Institute of Standards and Technology. (2026, November 4). CVE-2026-102895 detail. NVD. https://nvd.nist.gov/vuln/detail/CVE-2026-102895 ↩︎ ↩︎

  5. National Institute of Standards and Technology. (2026, November 6). CVE-2026-123218 detail. NVD. https://nvd.nist.gov/vuln/detail/CVE-2026-123218 ↩︎ ↩︎

  6. IBM. (2026, November 7). Security bulletin: IBM Aspera Faspex has addressed a vulnerability that could allow authorization bypass. IBM Support. https://www.ibm.com/support/pages/node/7295921 ↩︎ ↩︎