Versions of Puppet Enterprise prior to 2016.4.5 or 2017.2.1 failed to mark MCollective server private keys as sensitive (a feature added in Puppet 4.6), so key values could be logged and stored in PuppetDB. These releases use the sensitive data type to ensure this won't happen anymore.

Versions of Puppet prior to 4.10.1 will deserialize data off the wire (from the agent to the server, in this case) with a attacker-specified format. This could be used to force YAML deserialization in an unsafe manner, which would lead to remote code execution. This change constrains the format of data on the wire to PSON or safely decoded YAML.

Versions of MCollective prior to 2.10.4 deserialized YAML from agents without calling safe_load, allowing the potential for arbitrary code execution on the server. The fix for this is to call YAML.safe_load on input. This has been tested in all Puppet-supplied MCollective plugins, but there is a chance that third-party plugins could rely on this insecure behavior.

The mcollective-sshkey-security plugin before 0.5.1 for Puppet uses a server-specified identifier as part of a path where a file is written. A compromised server could use this to write a file to an arbitrary location on the client with the filename appended with the string "_pub.pem".

On Windows installations of the mcollective-puppet-agent plugin, version 1.12.0, a non-administrator user can create an executable that will be executed with administrator privileges on the next "mco puppet" run. Puppet Enterprise users are not affected. This is resolved in mcollective-puppet-agent 1.12.1.

The Puppet Communications Protocol in Puppet Enterprise 2015.3.x before 2015.3.3 does not properly validate certificates for the broker node, which allows remote non-whitelisted hosts to prevent runs from triggering via unspecified vectors.

MCollective 2.7.0 and 2.8.x before 2.8.9, as used in Puppet Enterprise, allows remote attackers to execute arbitrary code via vectors related to the mco ping command.

The Puppet Communications Protocol (PCP) Broker incorrectly validates message header sizes. An attacker could use this to crash the PCP Broker, preventing commands from being sent to agents. This is resolved in Puppet Enterprise 2016.4.3 and 2016.5.2.

The mcollective-puppet-agent plugin before 1.11.1 for Puppet allows remote attackers to execute arbitrary code via vectors involving the --server argument.

Open redirect vulnerability in the Console in Puppet Enterprise 2015.x and 2016.x before 2016.4.0 allows remote attackers to redirect users to arbitrary web sites and conduct phishing attacks via a // (slash slash) followed by a domain in the redirect parameter. NOTE: this vulnerability exists because of an incomplete fix for CVE-2015-6501.