CVE-2026-55653: OpenSSH DH-GEX Double Free Denial of Service Vulnerability

Technical summary

The vulnerability stems from improper memory management in OpenSSH's DH-GEX known-group validation routine when operating in FIPS mode. When a client receives and processes group parameters from an SSH server during the DH-GEX key exchange, a code path exists where the same memory region is freed twice under certain conditions. An attacker controlling a malicious SSH server can craft specific DH-GEX group parameters that trigger this double-free condition, causing heap corruption and immediate process termination. The flaw is classified as CWE-415 (Double Free). The attack surface is the SSH client itself; successful exploitation requires the user to connect to the attacker's server, making it user-dependent but not requiring network-level manipulation.

Business impact

For organizations with SSH-dependent workflows, this vulnerability introduces availability risk. End users attempting to connect to compromised or attacker-controlled SSH servers will experience client crashes, disrupting remote access sessions, CI/CD pipelines relying on SSH key exchange, and automation tasks. The impact is typically temporary and recoverable (users restart the client), but repeated exploitation could be weaponized to disrupt critical operations. Organizations leveraging OpenSSH in FIPS-compliant environments—such as those in regulated sectors (government, finance, healthcare)—face heightened risk because FIPS mode is often mandatory. The vulnerability does not compromise confidentiality or integrity of data, only availability.

Affected systems

The vulnerability affects OpenSSH implementations across OpenBSD and Red Hat ecosystems, including Red Hat Enterprise Linux (multiple versions), Red Hat OpenShift Container Platform, and Red Hat Hardened Images. Any system running an SSH client in FIPS mode is vulnerable when connecting to an untrusted or compromised SSH server. Particular attention should be paid to organizations deploying OpenSSH clients in containerized or cloud-native environments (OpenShift), where lateral movement or supply-chain compromise of SSH endpoints could expose multiple workloads simultaneously. Systems not in FIPS mode may have reduced exposure depending on OpenSSH version and configuration.

Exploitability

The vulnerability has a CVSS score of 4.3 (MEDIUM severity) with a network attack vector and low attack complexity, but requires user interaction—specifically, a user must initiate an SSH connection to the attacker's server. This requirement limits mass exploitation but does not eliminate practical risk. The attack is not complex to execute: an attacker needs only to set up a malicious SSH server and trick or socially engineer a user into connecting. No special authentication is required; the server can immediately trigger the flaw during the key exchange negotiation. The vulnerability is not listed on the CISA Known Exploited Vulnerabilities (KEV) catalog, suggesting active exploitation in the wild has not yet been documented, but the straightforward nature of the attack means exploitation proof-of-concept code could emerge quickly.

Remediation

Organizations should prioritize patching OpenSSH across all affected systems. Red Hat customers should apply security updates for RHEL, OpenShift Container Platform, and Hardened Images as they become available. OpenBSD users should update to the latest patched version of OpenSSH. For systems unable to patch immediately, mitigations include disabling FIPS mode if operationally feasible (though this may violate compliance requirements), restricting SSH client access to trusted servers only via firewall rules or SSH configuration (StrictHostKeyChecking, known_hosts management), and monitoring for unusual SSH client crashes in logs. User education emphasizing connection warnings and verification of server identities is also important.

Patch guidance

Wait for and apply official security patches from your vendor: Red Hat users should check their respective RHEL version advisories and OpenShift release notes; OpenBSD users should update OpenSSH via their package management system. Verify patch versions against the vendor advisory before deployment. Patches should be prioritized for systems operating in FIPS mode and for systems where SSH clients are heavily used (jump hosts, CI/CD systems, automation servers). After patching, validate that SSH functionality and FIPS compliance remain intact in your environment.

Detection guidance

Monitor system logs for unexpected SSH client process terminations, particularly in FIPS-enabled environments. Look for segmentation faults, abort signals, or heap corruption errors in syslog or application logs coinciding with SSH connection attempts. Network monitoring tools can identify connections to suspicious or unfamiliar SSH servers. Endpoint detection and response (EDR) solutions should flag abnormal process exits associated with ssh or sshd. Organizations using OpenSSH in CI/CD pipelines should track pipeline failures attributed to SSH client crashes and correlate them with connection attempts to unfamiliar hosts. Test your detection posture by simulating a malicious SSH server connection (in isolated environments only) and confirming your logging and alerting systems catch the event.

Why prioritize this

While the CVSS score is MEDIUM (4.3), prioritization should account for context: FIPS-compliance requirements make this vulnerability more severe than the base score suggests for regulated organizations. The user-interaction requirement and lack of current KEV listing lower urgency compared to unauthenticated remote code execution flaws, but the ease of exploitation and potential for supply-chain abuse (compromised SSH endpoints) warrant timely patching. Organizations running SSH-critical infrastructure should patch within 30 days; others can follow standard patch cycles.

Risk score, explained

The CVSS 3.1 score of 4.3 reflects: Network-accessible attack vector (AV:N) with low attack complexity (AC:L) and no privileges or special conditions required, but user interaction needed (UI:R). The impact is limited to availability (A:L), with no confidentiality or integrity loss (C:N, I:N). The user-interaction requirement prevents this from being a critical remote code execution flaw; however, the straightforward nature of the attack and FIPS-specific exposure elevate practical risk for compliance-bound organizations. The base score does not account for business context, making organizational risk assessment essential.

Frequently asked questions