cobbler_project Vulnerabilities

Improper Authorization in GitHub repository cobbler/cobbler prior to 3.3.2.

An issue was discovered in Cobbler before 3.3.1. Files in /etc/cobbler are world readable. Two of those files contain some sensitive information that can be exposed to a local user who has non-privileged access to the server. The users.digest file contains the sha2-512 digest of users in a Cobbler local installation. In the case of an easy-to-guess password, it's trivial to obtain the plaintext string. The settings.yaml file contains secrets such as the hashed default password.

An issue was discovered in Cobbler through 3.3.1. Routines in several files use the HTTP protocol instead of the more secure HTTPS.

An issue was discovered in Cobbler before 3.3.1. In the templar.py file, the function check_for_invalid_imports can allow Cheetah code to import Python modules via the "#from MODULE import" substring. (Only lines beginning with #import are blocked.)

Cobbler before 3.3.0 allows authorization bypass for modification of settings.

Cobbler before 3.3.0 allows arbitrary file write operations via upload_log_data.

Cobbler before 3.3.0 allows log poisoning, and resultant Remote Code Execution, via an XMLRPC method that logs to the logfile for template injection.

A flaw was found in cobbler software component version 2.6.11-1. It suffers from an invalid parameter validation vulnerability, leading the arbitrary file reading. The flaw is triggered by navigating to a vulnerable URL via cobbler-web on a default installation.

It was found that cobbler 2.6.x exposed all functions from its CobblerXMLRPCInterface class over XMLRPC. A remote, unauthenticated attacker could use this flaw to gain high privileges within cobbler, upload files to arbitrary location in the context of the daemon.

Cobbler version up to 2.8.2 is vulnerable to a command injection vulnerability in the "add repo" component resulting in arbitrary code execution as root user.

The set_mgmt_parameters function in item.py in cobbler before 2.2.2 allows context-dependent attackers to execute arbitrary code via vectors related to the use of the yaml.load function instead of the yaml.safe_load function, as demonstrated using Puppet.