Incus Container Image Templating willkürliches Lesen und Schreiben von Hostdateien

The impact of CVE-2026-23954 is severe. An attacker who can launch a container with a custom image, such as a member of the 'incus' group, can craft a malicious metadata.yaml file containing directory traversal sequences or symbolic links. This allows them to read arbitrary files from the host system, potentially exposing sensitive configuration data, credentials, or source code. More critically, the attacker can also write arbitrary files, effectively gaining control over the host environment and executing arbitrary commands. This could lead to complete system compromise, data exfiltration, or denial of service. The ability to exploit this vulnerability in IncusOS expands the potential attack surface to the underlying operating system.

Organizations utilizing Incus for container orchestration, particularly those with container users possessing elevated privileges or access to sensitive data, are at risk. Shared hosting environments where multiple users can launch containers with custom images are also particularly vulnerable. Environments using older, unpatched Incus versions are at the highest risk.

• linux / server:

journalctl -u incusd | grep -i "template parsing error"

• linux / server:

lsof -i :8080 | grep incusd # Check for Incus process listening on standard port

• generic web:

curl -I http://<incus_ip>:8080/api/v1/ | grep -i "server: incus"

1. 2026-01-22Disclosure

   disclosure

1. Reserviert2026-01-19
2. Veröffentlicht2026-01-22
3. Geändert2026-03-27
4. EPSS aktualisiert2026-03-23

The primary mitigation for CVE-2026-23954 is to upgrade Incus to version 6.1.1 or later, which includes the necessary fixes to prevent directory traversal and symbolic link exploitation. If upgrading immediately is not possible, consider restricting the ability of users to launch containers with custom images. Implement strict access controls to limit who can create and manage container images. While not a direct fix, consider using a Web Application Firewall (WAF) or proxy to filter potentially malicious requests targeting the image templating functionality. Monitor container logs for suspicious activity, particularly attempts to access or modify files outside of expected container directories.