This looks like a great idea but I have a few concerns/questions about it:<p>1) As others have mentioned, there needs to be transparency on what each script does to make sure that it is not malicious. Is there a way to inspect the code without running it?<p>2) How do you protect the communication between stackfu and the user's servers? I saw in the video that you provide some keys, but is the connection thoroughly secured to prevent any potential spoofing? What about a potential attack that would come through the site itself and potentially access all users' activated servers? I would be very cautious to have stackfu installed as a daemon on a production server without knowing it's secure inside out -- even if I only enable it when I need it.<p>3) Did you consider developing a stackfu shell client? Using the command line client, you could provide some basic features like searching and running scripts. Since the scripts are started from the server itself this might be more secure. Although, you still need to make sure that the scripts are actually coming from stackfu (in case it was somehow spoofed), similar to package distribution systems (yum, apt, etc.). Another utility for the client could be to allow/disallow queued scripts sent from stackfu and maybe even a monitoring tool (see the output, previous logs, etc.).<p>4) How do you handle different *nix configurations? Do script authors need to make different scripts for each distribution? Do you provide a facility to only search for scripts that are compatible with your server(s)?<p>5) Aside from the number of deployments, do you have a rating system? Script failure and success rates? The number of deployments is hardly enough for me to choose from say 20 different Wordpress installation scripts.