Laravel Reverb insecure deserialization enables remote code execution

Note: This data is for reference and cybersecurity research purposes only.CyStack advises users not to use this information for unlawful purposes.

What is CVE-2026-23524?

CVE-2026-23524 is a vulnerability classified as Deserialization of Untrusted Data, affecting reverb (affected versions: < 1.7.0). This vulnerability is rated Critical, with a CVSS score of 9.8. Current sources do not report this vulnerability as exploited.

Overview

Original source data

Laravel Reverb provides a real-time WebSocket communication backend for Laravel applications. In versions 1.6.3 and below, Reverb passes data from the Redis channel directly into PHP’s unserialize() function without restricting which classes can be instantiated, which leaves users vulnerable to Remote Code Execution. The exploitability of this vulnerability is increased because Redis servers are commonly deployed without authentication, but only affects Laravel Reverb when horizontal scaling is enabled (REVERB_SCALING_ENABLED=true). This issue has been fixed in version 1.7.0. As a workaround, require a strong password for Redis access and ensure the service is only accessible via a private network or local loopback, and/or set REVERB_SCALING_ENABLED=false to bypass the vulnerable logic entirely (if the environment uses only one Reverb node).

Affected products and scope

  • laravel/reverb: the normalized record marks versions < 1.7.0 as affected. The advisory narrative specifically describes the affected line as 1.6.3 and below.
  • Exposure requires Laravel Reverb horizontal scaling with REVERB_SCALING_ENABLED=true, where Reverb servers communicate through Redis PubSub.
  • Confirmed patched release: 1.7.0.
  • A deployment with one Reverb node and REVERB_SCALING_ENABLED=false bypasses the vulnerable logic according to the advisory, but the dependency version should still be checked and upgraded where possible.

Technical details

Laravel Reverb uses Redis PubSub in its horizontal scaling path to relay events between Reverb nodes. In the affected implementation, PusherPubSubIncomingMessageHandler passed event fields obtained from Redis to PHP's unserialize() without restricting which classes could be instantiated. The patch updates src/Protocols/Pusher/PusherPubSubIncomingMessageHandler.php so that the relevant deserialization calls allow only Application, PendingMetric, and MetricType. This is CWE-502, where untrusted data is deserialized without sufficient class or type restrictions. Exploitation requires crafted data to reach the Redis channel of a Reverb instance running with REVERB_SCALING_ENABLED=true; the advisory states that unauthenticated Redis increases exposure. The available record does not specify the external path by which an attacker obtains Redis publish or write access, so reachability must be assessed for each deployment.

Exploitability

The vulnerability is network reachable and applies to Laravel Reverb's horizontal scaling logic when REVERB_SCALING_ENABLED=true. The supplied attack characteristics indicate that no privileges or user interaction are required and that attack complexity is low. Redis deployments without authentication increase risk when Redis can be reached or receives data from an unauthorized source. A multi-node deployment is not required for the vulnerable code to exist, but a single-node deployment with scaling disabled bypasses the vulnerable logic according to the advisory. The supplied record marks public exploit as false and known exploitation as null; the consulted sources do not establish observed exploitation.

Technical impact

If attacker-controlled Redis data reaches an affected Reverb instance, processing it can result in remote code execution under the Reverb process context. The record indicates that the technical outcome can affect confidentiality, integrity, and availability within the same security scope. The actual blast radius depends on the Reverb service account, Redis access controls, and host or container isolation. The vulnerability alone does not establish container escape, compromise of other hosts, or access to systems beyond the process's permissions.

Business impact

Successful exploitation could cause the Reverb process to execute attacker-controlled code. Depending on the service account and the isolation of the host or container, possible consequences include:

  • Reading data or secrets accessible to the Reverb process.
  • Modifying data, source files, or configuration within the process's permissions.
  • Disrupting WebSocket service or consuming host resources.
  • Incident response work to isolate systems and rotate credentials if compromise is suspected.

These are possible consequences of remote code execution, not evidence that a particular deployment has been breached.

Remediation

  1. Upgrade laravel/reverb to the confirmed fixed release 1.7.0. Verify that the upgraded version is the code actually deployed, not only the version declared in the source tree.
  2. If upgrading is not yet possible, require a strong Redis password and restrict Redis access to a private network or local loopback.
  3. If the environment uses only one Reverb node, set REVERB_SCALING_ENABLED=false to bypass the vulnerable logic entirely. Do not use this workaround where horizontal scaling is required.
  4. Recheck every application or Reverb node that uses the same Redis service, because updating one node does not demonstrate that other nodes are patched.
  5. If Redis was accessible from an untrusted network, review credentials, access controls, and related publish activity as a precaution.

Detection

  1. Inventory Composer dependencies and the deployed laravel/reverb version across every application, container, and Reverb host.
  2. Check the effective value of REVERB_SCALING_ENABLED, the number of Reverb nodes, and whether Redis PubSub is used to synchronize those nodes.
  3. Verify that Redis requires a strong password and review firewall rules, security groups, ACLs, or bindings to confirm that Redis is reachable only through a private network or local loopback.
  4. If code-level validation is needed, inspect PusherPubSubIncomingMessageHandler and confirm that deserialization calls restrict allowed classes as shown in the patch.
  5. Review Redis connections and publish activity from unexpected sources as a precautionary monitoring measure. The available evidence does not define a specific IOC or log signature, so the absence of suspicious logs does not prove that a deployment is safe.
Sources (20)
Learn more

Run an in-depth assessment with complete web risk management

CyStack VulnScan continuously discovers assets, validates vulnerabilities, and helps security teams prioritize remediation across the organization.