Exploitation reporting has changed the response to the OpenSSH password bypass known as SSHole, tracked as CVE-2026-123158. GreyNoise's November 6 summary says it is monitoring successful exploitation as well as scanning and attempts. Independently, Previdian's sensor page, reviewed November 17, records attempts from November 6 through November 16. For organizations that left a qualifying SSH service exposed, verifying a fix is now only half the job: investigate whether an earlier login gave someone access to the host and the credentials behind it.
This is a follow-up to Bulwark's November 5 disclosure coverage, which explicitly noted that its reviewed sources did not establish exploitation in the wild. The new element is exploitation evidence after that disclosure, not a new vulnerability or a newly released patch today.
What the exploitation evidence establishes
GreyNoise's first-party summary explicitly includes successful exploitation in the activity it is monitoring. That supports treating the issue as an exploited access path, rather than only a theoretical bypass with public demonstration code. The short summary does not establish how many production organizations were compromised, which actors were responsible, or what they did after entry.
Previdian provides a separate, narrower observation: its displayed snapshot contains 46 exploitation attempts from 27 source IP addresses across four sensors, with the last observation on November 16. The company expressly cautions that its telemetry does not prove the target was vulnerable or that an attempt succeeded. Those numbers must not be presented as breached-host counts, and source-IP geography is not national attribution.
The linked GreyNoise live-activity site could not be retrieved during this review. We therefore do not make detailed payload, persistence or credential-theft claims from that report. Neither the available summary nor the sensor counters establishes the full scale of compromise.
Scope the exposed systems, not every Linux machine
The original researcher, Özgün Kültekin of Wakefield, describes a conditional authentication failure involving OpenSSH on musl-based Linux without PAM. The affected path requires password authentication to remain enabled, a target account whose stored password field is exactly a single asterisk, and another account with a real password hash. Under those conditions, an oversized password can authenticate to the asterisk-marked account.
A user who normally signs in with a key is not necessarily protected if the server still permits the vulnerable password path. Conversely, an Alpine-based container without a listening SSH server is not exposed merely because it uses musl. The researcher identifies glibc and PAM authentication paths as outside this particular bypass.
Privilege also matters. The researcher identifies the default user on Alpine's official AWS and Azure images as having passwordless doas privileges. Where all vulnerable conditions are met, access to that account can therefore lead to root. That is a conditional escalation consequence, not evidence that every Alpine cloud instance has been breached.
Actions for small businesses and government contractors
- Establish the historical exposure window. Identify actual OpenSSH server processes on cloud hosts, appliances and custom images. Record the libc, PAM use, effective password policy, relevant account state and network reachability. Determine when fixed software or a configuration mitigation became effective; current settings alone do not reconstruct the earlier state.
- Close the entry path and verify the running fix. The researcher recommends OpenSSH 10.6p2 or disabling password authentication where it is unnecessary. OpenSSH dates Portable 10.6p2 to November 5. Use supplier-documented fixed packages or backports and verify the active daemon has consumed them. Check applicable configuration overrides, keep a tested console recovery route, and update reusable images as well as running hosts.
- Preserve and correlate authentication evidence. Retain available SSH, system, cloud and network logs before rebuilding. Investigate successful password logins to accounts intended to be key-only or unable to accept passwords. Compare timestamps with account changes, source addresses, session activity and administrator records. Treat an unexpected success as an investigation lead, not automatic proof of this specific exploit.
- Investigate the authority behind a suspicious session. Review unauthorized keys, new accounts, scheduled jobs, startup changes and unusual outbound connections using trusted response tooling. These are response checks, not behaviors established by the reviewed campaign evidence. If integrity cannot be established, rebuild from a verified image rather than assuming a patched daemon has removed existing access.
- Recover exposed credentials from a clean environment. Scope the deployment keys, cloud credentials, service tokens and administrative connections the affected account could read or use. Revoke and replace material justified by that exposure assessment, check downstream access, and validate that old credentials no longer work. Do not rotate sensitive replacements on a host whose integrity remains uncertain.
What a useful provider response should contain
If a hosting vendor or managed service provider owns the affected system, request evidence rather than a generic assurance that patching is complete: which serving package includes the fix, when it became active, whether the vulnerable authentication configuration existed, what historical logs remain, and whether unauthorized access was investigated. For contractor environments, map suspected access to the systems and data in scope and follow applicable contractual reporting procedures; this reporting does not establish that any particular customer's data was accessed.
Bulwark assessment
The practical change is from preventing a possible login to testing whether an exploited login path was available before remediation. The affected subset is narrower than all OpenSSH or all Linux, but the authority reachable from a single administrative host can be substantial. A sound closure record should separate three conclusions: the entry path is fixed, the earlier exposure was investigated, and any compromised host or credential trust was recovered. A patch receipt establishes only the first, and only after runtime verification.
Primary source: GreyNoise's November 6 SSHole monitoring summary.
Primary source: Previdian's first-party exploitation-attempt observations and telemetry limitations.
Primary source: Özgün Kültekin's original authentication-bypass research, affected conditions and mitigation guidance.
Primary source: OpenSSH release notes identifying Portable 10.6p2's November 5 release.

