I would like to see if there is a way to add conditional logic to the robots.txt file so that when we push from DEV to PRODUCTION and the robots.txt file is pushed, we don't have to remember to NOT push the robots.txt file OR edit it when it goes live.

My specific situation is this:

I have www.website.com, dev.website.com and new.website.com and somehow google has indexed the DEV.website.com and NEW.website.com and I'd like these to be removed from google's index as they are causing duplicate content.

Should I:

a) add 2 new GWT entries for DEV.website.com and NEW.website.com and VERIFY ownership - if I do this, then when the files are pushed to LIVE won't the files contain the VERIFY META CODE for the DEV version even though it's now LIVE? (hope that makes sense)

b) write a robots.txt file that specifies "DISALLOW: DEV.website.com/" is that possible? I have only seen examples of DISALLOW with a "/" in the beginning...

Hope this makes sense, can really use the help!  I'm on a Windows Server 2008 box running ColdFusion websites.