On this page (7 sections)

CVE-2026-124079 is a critical flaw in rConfig, a network configuration management tool built on Laravel. According to the NVD record, an unauthenticated attacker on the network can push a crafted serialized job onto the Redis-backed Laravel Horizon queue and obtain code execution on the rConfig server.

If you run rConfig Core 8.0.0 up to but not including 8.2.19, or rConfig Pro before 8.4.1, treat this as a priority. The record says an attacker can read database credentials, stored device credentials and API tokens, so a compromised server can become a stepping stone into the network equipment it manages.

What happened

The flaw was published on 2026-11-09. The record says affected deployments run the Laravel Horizon queue on Redis without authentication. Anyone who can reach that Redis instance can place a job on the queue, and the application will process it as if it were legitimate.

The research found little coverage beyond the NVD record and the pages it references, and no independent confirmation of exploitation or real-world impact. The record points to a vendor release for Core 8.2.19, the rConfig security advisories page and a VulnCheck advisory. Read the vendor pages for upgrade notes before you change anything.

How the weakness works

Deserialization turns stored data back into live objects. If the application does this with data an attacker controls, and does not limit which object types are allowed, the attacker can choose what gets created. That is the class of bug described by CWE-502.

Per the NVD description, Laravel's queue handler unserializes the job without class restrictions. This lets an attacker instantiate an existing application job class, TaskDownloadRun, and make it invoke an Artisan command of their choosing. The bundled laravel/tinker package then allows execution of arbitrary PHP and operating system commands as the user the queue worker runs as.

The record lists these consequences:

  • Arbitrary PHP and OS command execution as the queue worker's user.
  • A persistent web shell written into the web-served storage/app/public directory.
  • Exposure of database credentials, stored device credentials and API tokens.
This article does not include payload details. The point for defenders is that an unauthenticated Redis queue is effectively a remote code execution interface.

When Redis is not directly reachable

The record adds a second path. Where an attacker cannot reach Redis directly, an authenticated user can deliver the same payload through the application by configuring a device target that points at the rConfig server's loopback address. So limiting network access to Redis reduces the first path but does not remove the need to patch, and low-privilege accounts matter too.

Who is realistically exposed

  • Teams running rConfig Core 8.0.0 to 8.2.18, or rConfig Pro earlier than 8.4.1.
  • Deployments where Redis listens on a non-loopback interface, or is reachable from user networks, other servers or the internet, with no password set.
  • Installations where many people hold rConfig accounts, since an authenticated user can use the loopback route described above.
  • Servers where the queue worker runs as a privileged user, because code runs with that user's rights.

The research does not say how many installations are exposed or whether attacks are underway. Absence of reports is not evidence of safety, so check your own systems.

How to check your exposure

First find your installed version in the rConfig web interface or the release notes you deployed from, and compare it with the fixed versions above. Then check what Redis is listening on. The command below is read-only and lists listening TCP sockets for Redis on Linux. A listener on 127.0.0.1 only is far better than 0.0.0.0 or a LAN address.

List listening sockets for Redis (run on the rConfig server, as root)Shell
ss -ltnp | grep -i redis

Next, look for files that should not be in the public storage directory. This read-only command lists PHP files under storage/app/public, where legitimate uploads are not normally PHP. Replace the path with your rConfig install directory. Any result deserves investigation.

Find PHP files in the web-served storage directory (run on the rConfig server)Shell
find /path/to/rconfig/storage/app/public -type f -name "*.php" -ls

This finds only one kind of artifact. A web shell can use other names or locations, so also review web server access logs for unfamiliar requests to files in that directory.

What to do now

  1. Back up the rConfig server and database, then upgrade rConfig Core to 8.2.19 or later, or rConfig Pro to 8.4.1 or later, following the vendor's instructions.
  2. Restrict Redis to the local host or a private interface, and block its port (6379 by default) at the host firewall and network edge for everything except the rConfig server.
  3. Enable Redis authentication if it is not already enabled, and confirm rConfig is configured with the matching credentials.
  4. Check storage/app/public for unexpected files, and review web and system logs for activity you cannot explain.
  5. If you find signs of compromise, rotate database passwords, stored device credentials and API tokens, since the record says these can be read.
  6. Review who holds rConfig accounts and remove any that are not needed.
  7. Run the queue worker as a dedicated low-privilege user, not root.
Caution. Upgrading and changing Redis settings can interrupt scheduled jobs and configuration backups. Do this in a maintenance window, and make sure rConfig can still authenticate to Redis afterwards.

If you suspect the server was already abused, follow a full incident process rather than only patching. Our guide on what to do when a website is hacked covers the general steps.

Detection and monitoring

  • Alert on any connection to Redis from hosts other than the rConfig server.
  • Watch for new or modified files in storage/app/public and for PHP execution from that path.
  • Monitor queue activity for jobs you did not schedule, and for unexpected child processes spawned by the queue worker.
  • Watch for outbound connections from the rConfig server to unknown destinations, and for logins to managed devices from unusual sources.

Exposed services like Redis are easy to miss until someone looks from the outside. Webforta Scan runs read-only scans of sites you verify, and Webforta Monitor can track certificate, DNS and header changes. Neither replaces patching or network controls for an internal tool like this, but they help you notice when something public changes.

For broader background, see our small business website security guide.

Timeline

Known dates from the research
DateEvent
2026-11-09CVE-2026-124079 published; fixed versions rConfig Core 8.2.19 and Pro 8.4.1 referenced

No exploitation reports or earlier disclosure dates appear in the research. We will not guess at them.

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-2026-124079: rConfig Core 8.0.0 before 8.2.19 and rConfig Pro before 8.4.1 contain a deserialization of untrusted data vulnerability that allows unauthenticated network…nvd.nist.gov ·
  • rConfig
  • Redis
  • Laravel
  • RCE
  • deserialization
  • network management