RabbitMQ installers on Windows prior to version 3.8.16 do not harden plugin directory permissions, potentially allowing attackers with sufficient local filesystem permissions to add arbitrary plugins.

The official rabbitmq docker images before 3.7.13-beta.1-management-alpine (Alpine specific) contain a blank password for a root user. System using the rabbitmq docker container deployed by affected versions of the docker image may allow a remote attacker to achieve root access with a blank password.

RabbitMQ versions 3.8.x prior to 3.8.7 are prone to a Windows-specific binary planting security vulnerability that allows for arbitrary code execution. An attacker with write privileges to the RabbitMQ installation directory and local access on Windows could carry out a local binary hijacking (planting) attack and execute arbitrary code.

Pivotal RabbitMQ, versions 3.7.x prior to 3.7.21 and 3.8.x prior to 3.8.1, and RabbitMQ for Pivotal Platform, 1.16.x versions prior to 1.16.7 and 1.17.x versions prior to 1.17.4, contain a web management plugin that is vulnerable to a denial of service attack. The "X-Reason" HTTP Header can be leveraged to insert a malicious Erlang format string that will expand and consume the heap, resulting in the server crashing.

Pivotal RabbitMQ, 3.7 versions prior to v3.7.20 and 3.8 version prior to v3.8.1, and RabbitMQ for PCF, 1.16.x versions prior to 1.16.7 and 1.17.x versions prior to 1.17.4, contain two endpoints, federation and shovel, which do not properly sanitize user input. A remote authenticated malicious user with administrative access could craft a cross site scripting attack via the vhost or node name fields that could grant access to virtual hosts and policy management information.

Pivotal RabbitMQ, versions prior to v3.7.18, and RabbitMQ for PCF, versions 1.15.x prior to 1.15.13, versions 1.16.x prior to 1.16.6, and versions 1.17.x prior to 1.17.3, contain two components, the virtual host limits page, and the federation management UI, which do not properly sanitize user input. A remote authenticated malicious user with administrative access could craft a cross site scripting attack that would gain access to virtual hosts and policy management information.

CF CLI version prior to v6.45.0 (bosh release version 1.16.0) writes the client id and secret to its config file when the user authenticates with --client-credentials flag. A local authenticated malicious user with access to the CF CLI config file can act as that client, who is the owner of the leaked credentials.

An issue was discovered in amqp_handle_input in amqp_connection.c in rabbitmq-c 0.9.0. There is an integer overflow that leads to heap memory corruption in the handling of CONNECTION_STATE_HEADER. A rogue server could return a malicious frame header that leads to a smaller target_size value than needed. This condition is then carried on to a memcpy function that copies too much data into a heap buffer.

Pivotal RabbitMQ for PCF, all versions, uses a deterministically generated cookie that is shared between all machines when configured in a multi-tenant cluster. A remote attacker who can gain information about the network topology can guess this cookie and, if they have access to the right ports on any server in the MQ cluster can use this cookie to gain full control over the entire cluster.

Pivotal Spring AMQP, 1.x versions prior to 1.7.10 and 2.x versions prior to 2.0.6, expose a man-in-the-middle vulnerability due to lack of hostname validation. A malicious user that has the ability to intercept traffic would be able to view data in transit.