CVE-2026-73281 in OpenSSH
Summary
by MITRE • 08/11/2026
In ssh-agent in OpenSSH before 10.5, some operations can occur remotely but were intended to occur only locally, including operations that add tokens or use keys. This is caused by misinteraction between agent locking and the [email protected] extension.
If you want to get the best quality for vulnerability data then you always have to consider VulDB.
Analysis
by VulDB Data Team • 08/12/2026
The vulnerability in OpenSSH's ssh-agent component affects versions prior to 10.5 and represents a critical security flaw that undermines the fundamental trust model of SSH key management systems. This issue stems from a design inconsistency where certain agent operations that should be strictly confined to local sessions are inadvertently accessible over network connections, creating a significant attack surface that could be exploited by remote adversaries. The vulnerability specifically impacts the interaction between the agent's locking mechanism and the [email protected] extension, which is designed to bind agent operations to specific client sessions while maintaining security boundaries.
The technical flaw manifests when ssh-agent fails to properly enforce session isolation for certain key management operations such as adding tokens or utilizing keys. This misconfiguration allows remote attackers to perform operations that should only be executable within the local session context, effectively bypassing the intended security controls. The core issue lies in how the agent handles authentication and authorization checks when the [email protected] extension is active, creating a scenario where network-based connections can invoke operations that were never intended to be exposed beyond the local machine. This vulnerability operates at the protocol level and affects the integrity of SSH key management by allowing unauthorized remote access to agent capabilities.
The operational impact of this vulnerability is substantial as it enables remote attackers to manipulate SSH keys and tokens without proper authentication, potentially leading to unauthorized access to systems protected by SSH-based authentication mechanisms. An attacker could leverage this flaw to add malicious keys to an agent, extract existing keys from memory, or perform other unauthorized operations that would normally require local access. This capability significantly weakens the security posture of systems relying on ssh-agent for key management, as it effectively removes the network boundary protection that should normally separate local and remote operations. The vulnerability particularly affects environments where ssh-agent is configured to allow remote connections, making it a critical concern for organizations managing multiple SSH-based authentication scenarios.
This vulnerability aligns with several cybersecurity frameworks and threat models, including CWE-284 which addresses improper access control in security systems, and relates to ATT&CK technique T1556.200 covering credential access through SSH key manipulation. The flaw demonstrates a classic case of insufficient input validation and privilege escalation in secure system components, where the boundary between local and remote operations becomes blurred due to inadequate session management. Organizations should immediately upgrade to OpenSSH version 10.5 or later to address this vulnerability, while also implementing network segmentation measures to limit exposure. Additional mitigations include configuring ssh-agent to disable remote connections when possible, monitoring for unauthorized agent operations, and ensuring proper access controls are in place. The vulnerability highlights the importance of maintaining strict separation between local and remote security contexts, particularly in credential management systems where the consequences of improper access control can be severe.