SB2025102151 - Multiple vulnerabilities in LXD
Published: October 21, 2025
Breakdown by Severity
- Low
- Medium
- High
- Critical
Description
This security bulletin contains information about 7 secuirty vulnerabilities.
1) Information disclosure (CVE-ID: CVE-2025-54291)
The vulnerability allows a remote attacker to gain access to potentially sensitive information.
The vulnerability exists within the /1.0/images endpoint that returns different responses depending on the existence of the project. A remote non-authenticated attacker can enumerate existing projects.
2) Path traversal (CVE-ID: CVE-2025-54293)
The vulnerability allows a remote user to perform directory traversal attacks.
The vulnerability exists due to input validation error when processing directory traversal sequences within the validLogFileName() function in lxd/instance_logs.go. A remote user can send a specially crafted HTTP request and read arbitrary files on the system.
3) Information disclosure (CVE-ID: CVE-2025-54290)
The vulnerability allows a remote attacker to gain access to potentially sensitive information.
The vulnerability exists due to the images export API return different error messages depending on the project presence. A remote non-authenticated attacker can enumerate existing projects.
4) Information disclosure (CVE-ID: CVE-2025-54289)
The vulnerability allows a remote user to gain access to potentially sensitive information.
The vulnerability exists due to excessive data output by the API. A remote user with read-only permissions can obtain secret values necessary for WebSocket connections via an API call and execute arbitrary commands inside instances with the victim's privileges.
5) Spoofing attack (CVE-ID: CVE-2025-54288)
The vulnerability allows a remote user to perform spoofing attack.
The vulnerability exists due to the source container identification process in devLXD server uses process cmdline (command line) information when displaying process names. A remote user can use the "exec -a" command to spoof process name and impersonate another container.
6) Template injection (CVE-ID: CVE-2025-54287)
The vulnerability allows a remote user to read arbitrary files on the system.
The vulnerability exists due to improper input validation within the snapshot creation functionality of the Pongo2 template engine in LXD-UI. A remote user with permissions to modify instance settings can pass specially crafted input to the application and read arbitrary files on the system.
Payload example:
{% filter urlencode|slice:":100" %}{% include "/etc/passwd" %}{%endfilter %}
7) Cross-site request forgery (CVE-ID: CVE-2025-54286)
The vulnerability allows a remote attacker to perform cross-site request forgery attacks.
The vulnerability exists due to insufficient validation of the HTTP request origin when using client certificate authentication with LXD-UI. A remote attacker can trick the victim to visit a specially crafted web page and perform arbitrary actions on behalf of the victim on the vulnerable website.
Remediation
Install update from vendor's website.
References
- https://github.com/advisories/GHSA-xch9-h8qw-85c7
- https://github.com/advisories/GHSA-472f-vmf2-pr3h
- https://github.com/advisories/GHSA-p3x5-mvmp-5f35
- https://github.com/advisories/GHSA-3g72-chj4-2228
- https://github.com/advisories/GHSA-7232-97c6-j525
- https://github.com/advisories/GHSA-w2hg-2v4p-vmh6
- https://github.com/advisories/GHSA-p8hw-rfjg-689h