I’m designing a custom task that could contain confidential information, and I am trying to get a handle on the risk.
For instance, if I design a task that has a query parameter where the user enters the confidential info, where is that information exposed?
bigfix client log
bigfix console (the old info entered is displayed in the query parameter)
fixlet available from the webserver
where else??
Same task, but this time it gets the information/file from a null share (and the information is only there during the period of the deployment, and then deleted.)
bigfix client log (potentially, depending on how the information is pulled from the share/file)
where else?? shouldn’t be in the fixlet available from the web server itself right?
I don’t recommend that you do this… One day, we plan on adding a secured point-to-point channel to send secure info (such as passwords) to the agent in a way that would handle all the security concerns with man-in-the-middle attacks, replay attacks, etc. As it stands now, all data you send to the agents is secured (digital signatures) but not secret (it is available at the places you mention and in the database).
Your null session share has the same problem that makes the info accessible to anyone that knows enough to get it when it is online, but you are correct that BigFix will not record the contents of the file anywhere (unless, like you say, you use relevance substitution in an action, in which case the data will be in the agent logs until it rolls over).
Thanks Ben! As long as we understand the risks, we can make an informed choice about whether to accept that risk for the benefit, and also come up with ways to reduce the risk perhaps as well. The task I am looking at would involve spending about 5-10 minutes per computer setting up an agent, where deploying with a fixlet allows us to do all of them in about 10 minutes.
I am looking forward to any secure communication changes you make to the Bigfix product! That would solve several problems that I’ve come across over the years.