Back when I was studying for my RHCE, I needed the ability to practice trouble-shooting a Linux system without knowing how it was messed up. I quickly wrote some simple scripts to randomly mess up my test machine. In later discussions with some sysadmin friends of mine, I discovered that more people than me needed this tool. I took a week to redesign it, a week to re-implement it, and trouble-maker was born.
I designed the system to be as system-neutral as possible, but not having a whole lot of time, I decided to only support Red Hat Enterprise Linux, Fedora, and Novell SUSE Enterprise Linux. After setting up the project, I started looking for people to check the system on other operating systems. I was shocked at how many people wanted the other operating systems. As things stand, I have a basic system that is highly modular and multi-platform. If I get more professionally involved with doing Linux admin training, I will get back into writing modules.
However, in the time since I wrote this system, technology has changed. It would be more effective to take base SLES and RHEL virtual images and have a system that applies deltas to the image and reboots them. That would be more controlled and allow the system to be more consistent.