The idea behind robots.txt is to keep honest robots out of certain areas of a web site. One file not to include in robots.txt is…robots.txt. The bots need to read robots.txt to find out their restrictions. A robots.txt is assumed to exist in the root directory of a web site.
Another file to leave out of robots.txt is the sitemap file: generally sitemap.xml and/or sitemap.xml.gz. It is imperative that the bots find this file so they can pare down the amount of indexing needed.
Other files not to exclude from bots are verification files: the ones created to prove you have access to the web site. Some common applications are Google sitemap verification, McAfee Site Advisor, and Yahoo Site Explorer.
Doug