Abstract visualization of an internal server network with glowing nodes being scanned

Teams running Alpine Linux servers can now check their own networks for exposure to SSHHole (CVE-2026-123158), the OpenSSH flaw that affects releases older than 10.6p2 built against musl without PAM (Pluggable Authentication Modules): runZero’s SSH exposure scanner SSHamble added a “vuln-sshhole” check in version v0.5.1, released on November 5, 2026. Spanish security blog OpenSecurity has published a walkthrough on checking servers for SSHHole with SSHamble (in Spanish). This piece sticks to the operational side: getting the binary, scanning only what you own, turning the check into a recurring inventory job, and handling a host that comes back flagged.

The key points of checking for SSHHole with SSHamble in 30 seconds

  • SSHHole hits OpenSSH before 10.6p2 on musl without PAM, on accounts whose password field holds “*”.
  • Alpine has shipped fixes for every branch from 3.16 to edge; until a host is patched, PasswordAuthentication no is the mitigation.
  • runZero’s SSHamble v0.5.1 enables the vuln-sshhole check by default and writes results as JSONL.
  • Scans generate real login attempts that can trip fail2ban or other blocking, so warn the team and allowlist the scanner’s address.

SSHamble has been around for a while. HD Moore and Rob King introduced it at Black Hat USA 2024 in a talk called “Secure Shell in Shambles”, and runZero publishes it under the 2-clause BSD license. The SSHHole check arrived through pull request #53, which HD Moore merged on November 5, 2026. The first commits in that PR were authored by Paul Asadoorian and carry a “Co-Authored-By: Claude Opus 5.5” trailer.

One ground rule before anything else: point the tool only at systems you own, or at systems you have written permission to test. To whoever runs the target, a scanner trying to log in over SSH looks exactly like an attack.

Getting SSHamble and running a first scan

The code lives in the runZeroInc/sshamble repository on GitHub, and release v0.5.1 is the one that ships prebuilt binaries with the new check. Grab the build that matches the operating system and CPU architecture of the machine you will scan from, unpack it if it comes in an archive, and drop it somewhere on the path:

sudo install -m 0755 sshamble /usr/local/bin/sshamble

Scans run through the scan subcommand. Four options matter here: -u or –users for the accounts to try, -p or –ports for the ports, –checks to pick which checks run, and -o or –output for the results file. A first pass against a lab subnet looks like this:

sshamble scan -u root,alpine,nobody -p 22 --checks vuln-sshhole -o results.jsonl 192.168.1.0/24

Usernames are comma-separated with no spaces. The example tries root, alpine and nobody; adjust the list to the accounts that actually exist in your images. The vuln-sshhole check is already on by default, but naming it in –checks keeps the run focused on the one thing you are looking for. If any sshd listens on something other than port 22, reflect that in -p.

Output is JSONL (JSON Lines, one JSON object per line), which is easy to filter with jq, load into a spreadsheet or ship to a SIEM (security information and event management) platform. Start with a small segment, read the results by hand, and only then widen the scope.

A scheduled inventory that won’t lock out your scanner

A one-off scan answers today’s question, but new containers and freshly provisioned machines keep showing up. That makes this a good candidate for a recurring job, much like the APT patching checklist for Debian and Ubuntu fleets published earlier this month.

Keep in mind that the scan makes genuine authentication attempts. They land in each server’s logs and can trigger fail2ban, SSHGuard or whatever blocking is in place, which means the scanner gets banned or the on-call engineer gets paged. Tell the team first, always scan from a host with a fixed IP address, and allowlist that address. In fail2ban, that is the ignoreip setting in jail.local:

ignoreip = 127.0.0.1/8 ::1 10.20.0.15

For scheduling, a systemd timer is easier to keep an eye on than a crontab entry: runs show up in the journal and start times can be spread out. Start with a oneshot service at /etc/systemd/system/sshamble-sshhole.service:

[Unit]
Description=SSHHole inventory with SSHamble

[Service]
Type=oneshot
StateDirectory=sshamble
ExecStart=/usr/local/bin/sshamble scan -u root,alpine,nobody -p 22 --checks vuln-sshhole -o /var/lib/sshamble/sshhole.jsonl 10.20.0.0/24Code language: JavaScript (javascript)

Then add a matching timer at /etc/systemd/system/sshamble-sshhole.timer. This one fires weekly, outside business hours:

[Unit]
Description=Weekly SSHHole inventory

[Timer]
OnCalendar=Mon *-*-* 06:40:00
RandomizedDelaySec=20min
Persistent=true

[Install]
WantedBy=timers.targetCode language: JavaScript (javascript)

Enable it with systemctl daemon-reload followed by systemctl enable –now sshamble-sshhole.timer, and check the next run with systemctl list-timers. If cron is the house standard, a line starting with 40 6 * * 1 followed by the same command does the same job. Either way the results file is overwritten on every run, so keeping history means copying it elsewhere with a small wrapper script. It also helps not to sweep every range in the organization at once: splitting them across several timers keeps the load down and makes any alert easier to explain.

When a host comes back flagged

Mitigate first, then patch. The Alpine Linux advisory gives the stopgap: set PasswordAuthentication no in /etc/ssh/sshd_config and restart the service (on Alpine with OpenRC, rc-service sshd restart). Confirm that key-based login works before you do it, or you may lock yourself out.

The fix itself is available on Alpine from 3.16 through edge:

apk update && apk upgrade openssh

For containers, patching the running instance is not enough. Rebuild every Alpine-based image, for example with docker build –pull, and redeploy; otherwise the next restart brings the vulnerable package back. It is also worth reviewing accounts with no password or with “*” in /etc/shadow and deciding whether they need to exist at all:

awk -F: '$2 == "*" {print $1}' /etc/shadowCode language: JavaScript (javascript)

If the host was reachable from untrusted networks, go through its authentication logs for unexpected logins on those accounts. Historical activity data helps reconstruct what happened, and the guide to sar as your server’s flight recorder covers how to query it. Finish by rescanning the host and confirming it no longer shows up.

Frequently asked questions

Which servers are affected by SSHHole?

Servers running OpenSSH older than 10.6p2 built against musl without PAM, as on Alpine Linux. The flaw affects accounts with “*” in the password field.

Which SSHamble release includes the vuln-sshhole check?

Version v0.5.1, released on November 5, 2026 with prebuilt binaries on GitHub. The check is enabled by default and can be selected explicitly with –checks vuln-sshhole.

Can the scan get the scanning host blocked?

Yes. The login attempts are logged and can trigger fail2ban or similar tools, so warn the team and allowlist the scanner’s address beforehand.

How do you fix a vulnerable Alpine host?

Run apk update and apk upgrade openssh, and rebuild any Alpine-based container images. Until then, PasswordAuthentication no in sshd_config mitigates the issue.

Sources:

Scroll to Top