On this page (7 sections)

A high-severity vulnerability, tracked as CVE-2026-133500, affects Gladys Assistant. According to the GitHub Advisory Database entry, versions up to 4.71.0 contain an OS command injection flaw in the Backup Restore Endpoint component, and the advisory says an exploit is publicly available and might be used.

Anyone running Gladys Assistant should check which version they have and plan an upgrade to 4.72.0, which the advisory says mitigates the issue. Instances reachable over a network deserve the most urgency.

What happened

The record carries a published date of November 20, 2026. Its description says the flaw affects an unknown function in the file gateway.downloadBackup.js, part of the Backup Restore Endpoint component, and that manipulating the argument file_url leads to OS command injection. The advisory says the attack can be initiated remotely.

The advisory lists several references: the NVD entry, a GitHub issue (2494), the v4.72.0 release page and several VulDB pages. The research does not include the content of the issue or the release notes, so this article does not describe what they say beyond the advisory text.

Coverage is thin. The NVD entry repeats the same description as the GitHub advisory, so it is not independent confirmation. The GitHub advisory is marked unreviewed and lists affected and patched versions as unknown, while the description text itself names 4.71.0 and 4.72.0.

Technical details

In general, OS command injection happens when an application builds a system command from input a user controls and does not neutralize special characters first. The shell then runs more than the developer intended. Features that take a URL or file name and pass it to a helper program are a common place for this class of bug. The advisory identifies file_url as the manipulated argument here, but it does not explain the underlying code, so how the value reaches a command is not confirmed by the research.

The CVSS v4 metrics in the advisory describe a network attack vector, low attack complexity, no attack requirements and no user interaction. Privileges required are rated High. Impact is rated High for confidentiality, integrity and availability on both the vulnerable system and subsequent systems.

The High privileges rating may reduce the chance of a drive-by attack, but the description does not say what access is needed, so do not treat it as a reason to wait. Credentials are often reused or weak, and command injection can mean code execution on the host that runs the application.

Who is affected

  • Anyone running Gladys Assistant at version 4.71.0 or earlier.
  • Instances reachable from untrusted networks, such as through port forwarding, a reverse proxy or a tunnel, since the advisory says the attack can be initiated remotely.
  • Installations with weak or shared administrative credentials, because the CVSS metrics rate required privileges as High rather than none.
  • Hosts that store other data or sit near sensitive systems, because the impact ratings extend to subsequent systems.

The research does not describe how Gladys Assistant is typically deployed, so exposure depends on how you run it and who can reach it.

How to check your exposure

First find out which version you run, using the application's own interface or your installation records. If you run it in a container, this read-only command lists running containers with their image names and tags, so look for a Gladys image and compare its tag with 4.71.0:

List running containers and image tags (Docker hosts)Shell
docker ps --format '{{.Names}}\t{{.Image}}\t{{.Status}}'

Then check how the instance is reachable. If it answers on a public address, a forwarded port or a public hostname, treat it as internet-exposed. You can also search reverse proxy or web server logs for requests to backup-related paths that you did not make. This search is read-only, and the log path is an example you should adjust to your setup:

Search a web server access log for backup-related requests (adjust the log path)Shell
grep -i backup /var/log/nginx/access.log | tail -n 50

The research lists no specific indicators of compromise, so finding nothing does not prove a clean system.

What to do now

  1. Confirm the Gladys Assistant version you run. Anything at 4.71.0 or lower is affected.
  2. Upgrade to version 4.72.0 or later, following the project's normal update process and the v4.72.0 release page.
  3. Back up your Gladys data before upgrading and store the copy somewhere the application cannot write to.
  4. If you do not need internet access to the instance, remove it, for example by placing it behind a VPN or an authenticated reverse proxy (general hardening advice, not vendor guidance).
  5. Rotate administrator passwords and any API keys or tokens stored in or used by Gladys, especially if the instance was exposed.
  6. Limit the number of accounts with administrative rights and use unique, strong passwords.
  7. Review logs and running processes on the host for anything you do not recognize, and rebuild the host if you find signs of compromise.
Caution. Upgrading restarts the application and may briefly interrupt it. Make a backup first and choose a quiet time.
Tip. If you cannot upgrade immediately, restricting network access to the instance to trusted addresses is a reasonable general precaution. The research does not describe a vendor workaround, so this is not official guidance and upgrading remains the fix the advisory names.

Detection and monitoring

After upgrading, keep watching the host. Command injection can leave traces at the operating system level, so look for unexpected child processes of the Gladys service, new scheduled tasks, new user accounts and outbound connections you cannot explain. Keep access logs from any reverse proxy for several weeks.

If the instance is published under a public hostname, certificate and DNS changes can reveal tampering or forgotten exposure. Webforta Monitor tracks uptime, certificates and DNS for sites you own, and the free SSL checker reviews a certificate on demand. For guidance on responding to a possible intrusion, see website hacked: what to do.

What is still unclear

  • Whether the flaw is being exploited in the wild. The advisory says the exploit is public and might be used, but the research reports no confirmed attacks.
  • Which role or privilege level is needed. The CVSS metrics say High, but the description does not explain it.
  • Whether all versions before 4.71.0 are affected. The description says up to 4.71.0, while the GitHub record lists affected versions as unknown.
  • What the linked issue and release notes say in detail, since their content is not part of the research.

For more vulnerability coverage and practical guides, browse the Webforta blog and the security guides hub.

Sources

This article is based on the public sources below. Details can change as vendors and researchers publish more; check the original advisories before acting on version numbers.

  1. CVE-2026-133500 - GitHub Advisory Databasegithub.com
  2. CVE-2026-133500: A vulnerability was identified in Gladys Assistant up to 4.71.0.nvd.nist.gov ·
  • CVE-2026-133500
  • Gladys Assistant
  • command injection
  • patching