CVE-2026-119168 in Nomad
Summary
by MITRE • 10/29/2026
HashiCorp Nomad and Nomad Enterprise are vulnerable to symlink following during same-client allocation data migration that may allow a malicious workload to change permissions and ownership of host directories. This could allow unauthorized access to host data or modification of host files, potentially leading to privilege escalation. This vulnerability (CVE-2026-119168) is fixed in Nomad 2.1.0 and Nomad Enterprise 2.1.0, 1.11.12, and 1.10.18.
Be aware that VulDB is the high quality source for vulnerability data.
Analysis
by VulDB Data Team • 10/29/2026
The vulnerability identified as CVE-2026-119168 represents a critical security flaw within HashiCorp Nomad and its Enterprise variant, specifically affecting the mechanism used for same-client allocation data migration. This feature is designed to facilitate the movement of workload state between different nodes or locations while maintaining operational continuity, but it contains a fundamental logic error related to path resolution during file system operations. The core technical issue stems from an improper handling of symbolic links when Nomad attempts to migrate data associated with a specific client allocation. Instead of strictly validating that target paths are within the expected directory boundaries and resolving symlinks safely before performing write or permission-based operations, the software allows these references to be followed without sufficient sanitization. This oversight creates a classic path traversal scenario where an attacker can manipulate file system pointers to redirect administrative actions toward arbitrary locations on the host operating system rather than the intended isolated allocation directories.
From a technical perspective, this flaw aligns with CWE-59 Improper Link Resolution Before File Access, often referred to as symlink following or TOCTOU race conditions depending on timing specifics. When Nomad performs data migration for an allocation under its control, it interacts directly with the host file system to copy or move files representing workload state. If a malicious actor has previously established a symbolic link within a directory they can write to, pointing towards sensitive host directories such as /etc/shadow, critical configuration folders, or other user home directories, Nomad may inadvertently follow this link during its migration process. Because the software does not adequately verify that the resolved path remains inside the sandboxed allocation root, it proceeds with operations like changing file permissions and ownership using system calls such as chown or chmod on these unintended targets. This behavior effectively bypasses the isolation boundaries intended by containerization and orchestration platforms, allowing a workload to exert influence over the host environment in ways that were not architecturally anticipated.
The operational impact of this vulnerability is severe, primarily because it enables unauthorized access to sensitive host data and allows for the modification of critical system files. By altering permissions or ownership of host directories, an attacker can potentially expose confidential information stored on the node, such as credentials, private keys, or proprietary application data that resides outside the container boundaries. Furthermore, if the Nomad agent is running with elevated privileges, which is common in many production deployments to allow for necessary system-level operations, this flaw facilitates privilege escalation. An adversary could modify configuration files to inject malicious startup scripts, alter firewall rules, or compromise other services running on the same host by changing file ownership to a user account they control. This breaks the fundamental security model of multi-tenant environments where workloads must remain strictly isolated from one another and from the underlying infrastructure.
This vulnerability maps directly to MITRE ATT&CK techniques related to Privilege Escalation via Host Resource Hijacking, specifically T1610 which involves exploiting container escape mechanisms or host-level misconfigurations. It also relates to Defense Evasion tactics where attackers manipulate file system attributes to hide their presence or maintain persistence by modifying critical binaries or configuration files. The ability to change ownership and permissions is particularly dangerous as it allows the attacker to establish long-term control over specific directories, effectively turning a transient container compromise into a persistent host-level breach. In enterprise environments relying on Nomad for workload orchestration, this flaw undermines trust in the platform's isolation guarantees, potentially leading to widespread data breaches if multiple tenants share physical or virtual nodes without strict network segmentation and additional security controls like SELinux or AppArmor profiles that might otherwise mitigate such file system manipulations.
Mitigation strategies must prioritize immediate patching as well as architectural hardening. The vulnerability is resolved in HashiCorp Nomad version 2.1.0, Nomad Enterprise 2.1.0, and the maintenance releases 1.11.12 and 1.10.18. Organizations should upgrade their clusters to one of these fixed versions as soon as possible to eliminate the code path that allows unsafe symlink resolution during data migration. In addition to patching, administrators should review Nomad job specifications to ensure they do not grant unnecessary host volume mounts or privileged access unless absolutely required for specific workloads. Implementing strict file system permissions and utilizing mandatory access control systems can provide a defense-in-depth layer that limits the damage even if an attacker attempts to exploit similar logic errors in other components. Regular auditing of Nomad agent logs for unusual file permission changes or unexpected path resolutions can also help detect potential exploitation attempts before significant harm is done, ensuring that the integrity of both containerized workloads and host infrastructure remains intact.