See the following example from GitHub's gitignore repository: http://github.com/github/gitignore/raw/master/Node.gitignore
You can quickly recognize a url that will work to download an individual file vs downloading the html page. Look for the subdirectory /raw/ right before the branch name.