> ## Content Index
> Fetch the complete content index at: https://f4n6.co.uk/llms.txt
> Use this file to discover other available public pages before exploring further.

# SANS honeypots detect active scanning for sshole CVE-2026-123158
- URL: https://f4n6.co.uk/security-feed/sans-honeypots-detect-active-scanning-for-sshole-cve-2026-123158/
- Published: 2026-11-09T10:18:27.000Z
- Updated: 2026-11-09T10:18:27.000Z
- Author: Jeff Davies
- Tags: #security-feed, CVE-2026-123158

## 1\. Executive summary

SANS Internet Storm Center honeypots recorded a rise in SSH login attempts using unusually long passwords from 2026-11-01, and the first attempt matching the sshole exploit arrived on 2026-11-05, the same day OpenSSH 10.6p2 fixed the flaw. CVE-2026-123158 is scored CVSS 8.1 HIGH (CWE-394) in NVD and is not in CISA KEV; EPSS is 1%. Previdian reports confirmed root access on one of its Alpine Linux honeypots from a single source address, and says it has listed the bug as exploited while CISA has not. The flaw lets an attacker log in to passwordless accounts on musl-based OpenSSH builds with PAM disabled by sending a password longer than 256 characters, so any EMEA financial services estate running Alpine, OpenWrt, Buildroot or Yocto images with password SSH enabled is exposed to unauthenticated access. The client risk is high for those estates and low for estates on glibc OpenSSH with PAM, which the flaw does not affect.

## 2\. Regulatory framing

| Article                                                                     | Trigger (the fact in this item)                                                                                                                                | Practical impact                                                                                                                               |
| --------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------- |
| DORA Art. 17: ICT-related incident management process                       | Previdian reports confirmed successful root access on a honeypot via CVE-2026-123158, and SANS expects some of its honeypot logins succeeded                   | Firms running affected musl builds need a defined process to detect, contain and recover from unauthorised SSH access, not just a patch ticket |
| DORA Art. 24: digital operational resilience testing — general requirements | The exploit is a single-connection authentication bypass that testing should catch; Kültekin found it by throwing some 17,600 unusual login attempts at builds | Resilience testing should include oversized-password SSH attempts against passwordless accounts on any musl-based image in the estate          |
| NIS2 Art. 21(2)(d): supply chain security measures                          | The flaw is inherited through base images (Alpine's default openssh-server, Docker images, OpenWrt, Buildroot, Yocto) rather than through a firm's own code    | Supplier and image-provenance controls must confirm which musl-based images and appliances are in use and whether they carry the fix           |

## 3\. Technical analysis & attack chain

The vulnerability is an authentication bypass in OpenSSH on Linux systems that use the musl C library with PAM disabled. Accounts that should never accept a password (root, key-only deploy users, LDAP users, the default user on Alpine's AWS and Azure images) carry a "*" in place of a password hash in /etc/shadow. Since 2016, when OpenSSH changed sshd to borrow another account's salt and hash the password anyway for accounts with no usable hash (a fix for CVE-2016-6210), and since 2012, when musl's crypt() began refusing SHA-256 and SHA-512 passwords longer than 256 bytes and returning "*" on failure, the two behaviours combine: the "*" returned by musl matches the "*" stored for the passwordless account, and sshd lets the user in. Kültekin's summary is that such accounts "take any password longer than 256 characters". An attack needs password logins enabled (the OpenSSH default), an account with only "*" in its password field, and another local account that has a password hash. Root has shipped with a "*" since around Alpine 3.16, though OpenSSH's default PermitRootLogin keeps it out of reach unless an admin has changed it; SOCRadar argues indirect root access is the larger concern, because on Alpine's official AWS and Azure images the default alpine user is passwordless and has passwordless doas, which reaches root.

Confirmed steps observed in public reporting:

1. Scanner sends an SSH password longer than 256 characters. SANS saw the first exploit attempt on 2026-11-05 using an all "A" password.
2. Against an affected musl build with PAM disabled, sshd accepts the login to a passwordless account. Previdian confirmed successful root access on a honeypot running real Alpine Linux and OpenSSH.
3. The attacker runs reconnaissance through non-interactive SSH sessions. Previdian observed id, uname and hostname from source 89.116.89\[.\]18.

SANS's honeypots use Cowrie and collect only the first 255 characters of SSH passwords, so its view of the exploit payload is truncated. Its counts of long-password attempts were 23 on 2026-11-01, 5 on 2026-11-03, 58 on 2026-11-04, 232 on 2026-11-05, 713 on 2026-11-06, 246 on 2026-11-07 and 51 on 2026-11-08, with zero on each day from 2026-10-26 to 2026-10-31\. SANS states the 2026-11-01 passwords do not appear related to the exploit; it attributes them to buggy brute-force scripts that place post-login shell commands in the password field, and notes these may have worked. One such password string is recorded in §5\. SANS lists 18 source addresses among the initial scanners, with counts from 48 down to 2, spread across ASNs including COLT (AS8220), Datacamp (AS212238), B2 Net Solutions (AS55286), DMZHOST (AS48090), Uzbektelekom (AS202660), Telecom Armenia (AS12297), Korea Telecom (AS4766), LG DACOM (AS3786) and Scaleway (AS12876). SANS says it expects quite a few of the login attempts were successful on its honeypots and will follow up.

Previdian's account is single-sourced: it reports eight attempts, all from the single Japan-based address 89.116.89\[.\]18 and all caught by one sensor, with the first attempt less than nine hours after CVE publication. Previdian founder Ryan Dewhurst attributes the speed to AI and to the exploit being trivial. No source names a threat actor, and no MITRE ATT&CK profile is recorded for any actor in this item, so there is no attribution to confirm. The flaw was fixed in OpenSSH 10.6p2 on 2026-11-05\. Alpine Linux, OpenWrt, Buildroot and Yocto are named as musl-based systems to check, along with Docker images. The nickname sshole was given by Wakefield researcher Özgün Kültekin, who found the flaw.

## 4\. Mitigation & containment

P1 within 24h:

- Inventory every host and container image running OpenSSH on musl with PAM disabled. Alpine's default openssh-server package is built that way; check OpenWrt, Buildroot, Yocto and pulled Docker images.
- Upgrade OpenSSH to 10.6p2 or later on those systems. This is the fixed release named in public reporting.
- Where upgrade is not immediately possible, disable password authentication in sshd\_config and require keys, which removes the condition the exploit needs.
- Search SSH authentication logs for accepted logins to accounts that should never take a password (root, key-only deploy users, the default alpine user on AWS and Azure images) and for sessions that ran id, uname or hostname non-interactively.
- Block or alert on inbound SSH from the scanner addresses in §5 at the perimeter; treat them as untrusted.

P2 within 72h:

- Review PermitRootLogin settings. The OpenSSH default keeps root out of reach; any host where an admin changed it needs priority review.
- On Alpine AWS and Azure images, review the passwordless doas grant on the default alpine user, which reaches root.
- Rebuild container images from patched bases rather than patching running containers, and pin the base image to a version carrying OpenSSH 10.6p2.

P3 within 7 days:

- Add oversized-password SSH attempts (passwords over 256 characters) against passwordless accounts to resilience testing.
- Confirm which musl-based appliances and third-party images enter the estate and whether their suppliers ship the fix.

No vendor advisory beyond the OpenSSH 10.6p2 release is named in public reporting, and no CISA KEV remediation due date applies because the CVE is not in KEV.

## 5\. Indicators of compromise

| type | value                                                  | confidence                                                               | source            |
| ---- | ------------------------------------------------------ | ------------------------------------------------------------------------ | ----------------- |
| ipv4 | 89.116.89\[.\]18                                       | medium (single source, Previdian)                                        | Help Net Security |
| ipv4 | 156.238.252\[.\]37                                     | medium (SANS, buggy brute-force string, not confirmed as sshole exploit) | SANS ISC          |
| url  | hxxp://156\[.\]238\[.\]252\[.\]37:8000/bot\[.\]x86\_64 | medium (SANS, buggy brute-force string)                                  | SANS ISC          |
| ipv4 | 213.86.169\[.\]133                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 152.233.50\[.\]185                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 92.255.57\[.\]7                                        | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 195.178.110\[.\]78                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 109.94.173\[.\]13                                      | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 87.241.132\[.\]12                                      | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 52.119.124\[.\]122                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 185.2.148\[.\]137                                      | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 220.77.237\[.\]35                                      | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 82.74.132\[.\]129                                      | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 220.70.219\[.\]98                                      | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 154.80.128\[.\]173                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 121.190.39\[.\]238                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 121.163.224\[.\]97                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 58.151.184\[.\]226                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 62.210.173\[.\]100                                     | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 183.99.38\[.\]238                                      | low (SANS scanner list)                                                  | SANS ISC          |
| ipv4 | 112.162.23\[.\]132                                     | low (SANS scanner list)                                                  | SANS ISC          |

The scanner addresses are shared hosting, CDN and consumer ranges; several are legitimate providers, so blocking them wholesale carries collateral risk. The 156.238.252\[.\]37 string is a buggy brute-force artefact SANS says does not appear related to the sshole exploit; it is included because it is the only observed payload string and may indicate a host already compromised by other means.

```iocs
ipv4  89.116.89[.]18
ipv4  156.238.252[.]37
url  hxxp://156[.]238[.]252[.]37:8000/bot[.]x86_64
ipv4  213.86.169[.]133
ipv4  152.233.50[.]185
ipv4  92.255.57[.]7
ipv4  195.178.110[.]78
ipv4  109.94.173[.]13
ipv4  87.241.132[.]12
ipv4  52.119.124[.]122
ipv4  185.2.148[.]137
ipv4  220.77.237[.]35
ipv4  82.74.132[.]129
ipv4  220.70.219[.]98
ipv4  154.80.128[.]173
ipv4  121.190.39[.]238
ipv4  121.163.224[.]97
ipv4  58.151.184[.]226
ipv4  62.210.173[.]100
ipv4  183.99.38[.]238
ipv4  112.162.23[.]132

```

Behavioural indicators:

| behaviour                                                 | where to observe                       | confidence                                         |
| --------------------------------------------------------- | -------------------------------------- | -------------------------------------------------- |
| SSH password longer than 256 characters                   | sshd auth logs, Cowrie session logs    | high (mechanism confirmed by finder and Previdian) |
| Successful login to an account with "\*" in /etc/shadow   | sshd auth logs, /etc/shadow review     | high                                               |
| Non-interactive SSH sessions running id, uname, hostname  | process accounting, EDR, shell history | medium (single source, Previdian)                  |
| Post-login shell commands submitted in the password field | sshd auth logs                         | medium (SANS)                                      |

## 6\. Detection

The sources give no file artefacts, hashes or command-line flags specific to the sshole exploit itself, so a YARA rule would have nothing genuine to match. Detection rests on SSH authentication behaviour.

```yaml
title: SSH Login With Oversized Password Against Passwordless Account
id: 7f3c1a2e-sshole-2026-123158
status: experimental
description: Detects SSH authentication attempts using passwords longer than 256 characters, the condition that bypasses authentication on musl-based OpenSSH with PAM disabled (CVE-2026-123158).
author: Adverse Trace
date: 2026-11-09
references:

  - https://www.helpnetsecurity.com/2026/11/06/openssh-sshole-cve-2026-123158-exploitation/
  - https://isc.sans.edu/diary/rss/33498
logsource:
  product: linux
  service: sshd
detection:
  selection:
    message|contains: 'Failed password'
  longpass:
    message|re: 'password: .{257,}'
  condition: selection and longpass
fields:

  - src_ip
  - user
falsepositives:

  - Legitimate use of very long passphrases, which is rare
level: high

```

A second rule should alert on any accepted login to an account whose /etc/shadow entry is "\*", which is the successful-exploit condition; that requires joining sshd auth logs to a shadow-file inventory rather than a log-only rule.

## CVE assessment

1 referenced CVE

| CVE                                                                                | CVSS     | Exploited | EPSS | Summary                                                                                                                          |
| ---------------------------------------------------------------------------------- | -------- | --------- | ---- | -------------------------------------------------------------------------------------------------------------------------------- |
| [CVE-2026-123158](https://nvd.nist.gov/vuln/detail/CVE-2026-123158?ref=f4n6.co.uk) | 8.1 High | —         | 1%   | sshd in OpenSSH before 10.6p2 on Linux, when musl libc is used without PAM, lacks a workaround for a password authentication by… |

## 7\. Sources

- SANS Internet Storm Center, Johannes B. Ullrich, "Scans for Sshole Vulnerability (CVE-2026-123158)", https://isc.sans.edu/diary/rss/33498, 2026-11-08
- Help Net Security, "OpenSSH flaw exploited in the wild just hours after disclosure (CVE-2026-123158)", https://www.helpnetsecurity.com/2026/11/06/openssh-sshole-cve-2026-123158-exploitation/
- The Hacker News, "Hackers Target OpenSSH SSHole Flaw Within Hours of Disclosure, Gain Root on Honeypot", https://thehackernews.com/2026/11/hackers-exploited-openssh-sshole-flaw.html

## 8\. Adverse Trace position

Adverse Trace assesses CVE-2026-123158 as CVSS 8.1 HIGH, not in CISA KEV, with EPSS at 1%, and treats the exploitation as real but limited in what public reporting shows: SANS has scanner counts and one truncated payload string, and Previdian's confirmed root access is single-sourced, so the Previdian indicators should be verified before enforcement. The client impact is high for any EMEA financial services firm running Alpine, OpenWrt, Buildroot or Yocto images with password SSH enabled, because the bypass needs one connection and no password; estates on glibc OpenSSH with PAM are not affected. We will track SANS's follow-up on which honeypot logins succeeded, watch for further exploitation reports and any CISA KEV addition, and update this advisory if either changes the picture.

---

[Read the original source →](https://isc.sans.edu/diary/rss/33498?ref=f4n6.co.uk)

*Published via PulseTrace — Adverse Trace threat intelligence.*