CVE-2026-86335 in LXD
Summary
by MITRE • 09/28/2026
Missing Authorization in imageDownload in Canonical LXD before 5.0.10, 5.21.8, and 6.10 on Linux allows a project-restricted client to access private images from other projects via local fingerprint reuse during image or instance import requests.
VulDB is the best source for vulnerability data and more expert information about this specific topic.
Analysis
by VulDB Data Team • 09/28/2026
The vulnerability identified in Canonical LXD versions prior to 5.0.10, 5.21.8, and 6.10 represents a critical failure in access control mechanisms within the container management platform. This flaw specifically affects the imageDownload functionality, allowing project-restricted clients to bypass intended isolation boundaries. In a properly secured LXD environment, projects are designed to act as isolated units with distinct resource quotas and permission sets. However, due to this missing authorization check, an attacker operating under a restricted project context can exploit local fingerprint reuse during import operations to retrieve private images belonging to other projects. This undermines the fundamental security model of multi-tenancy that LXD relies upon for secure deployment in shared infrastructure environments.
From a technical perspective, the root cause lies in how the server validates requests against image metadata and access permissions. When an instance or image is imported, the system typically verifies whether the requesting client has permission to access the target resource based on its project membership and assigned roles. In this vulnerable scenario, the validation logic fails to strictly enforce these checks when handling local fingerprint references. By reusing a known valid fingerprint from another project's private image repository, an authenticated but restricted user can trick the server into serving that content without verifying if the current session has explicit authorization for it. This behavior effectively treats all accessible fingerprints as public or shared resources within the scope of any authenticated connection, regardless of their actual privacy settings.
The operational impact of this vulnerability is significant for organizations utilizing LXD in multi-project setups. It enables unauthorized data exfiltration where sensitive container images containing proprietary code, licensed software, or confidential configurations can be downloaded by users who should not have access to them. This breach of confidentiality compromises the integrity of the deployment pipeline and violates compliance requirements related to data isolation. Furthermore, if these private images contain vulnerabilities themselves, an attacker could potentially use this mechanism to study attack vectors against other projects within the same LXD instance. The ability to enumerate or download arbitrary private resources also aids in reconnaissance activities, providing attackers with valuable intelligence about the infrastructure's composition and potential weaknesses.
This issue aligns closely with CWE-284, which describes Improper Access Control, specifically highlighting failures where access control decisions are not enforced correctly during resource retrieval operations. In terms of offensive security frameworks, this vulnerability facilitates actions consistent with ATT&CK technique T1005, Data from Local System Retrieval, as it allows an actor to gather data stored locally on the host system without proper authorization. It also relates to T1530, Data from Cloud Storage Object, if interpreted broadly within cloud-native container contexts where images are treated as storage objects. The exploitation requires local access and authentication, placing it in the initial access or privilege escalation phases depending on the attacker's starting position relative to the LXD daemon.
Mitigation strategies must prioritize immediate patching of all affected LXD installations to versions 5.0.10, 5.21.8, or later where this authorization check has been corrected. Administrators should verify that their deployment models strictly adhere to least privilege principles, ensuring that project members only have access to resources explicitly assigned to them. Additionally, implementing robust monitoring and logging for image import operations can help detect anomalous behavior indicative of exploitation attempts. Regular audits of user permissions and project configurations are essential to maintain the integrity of isolation boundaries. Until patches are applied, restricting network exposure of the LXD socket and enforcing strict authentication mechanisms can reduce the attack surface available to potential adversaries seeking to exploit this missing authorization flaw.