Is this server in a domain environment? If so, you probably want to leave a lot of those predefined rules in place. To start with, I suggest only adding Allow rules for your HTTP and RDP traffic, and leave the rest as defaults.
If you really want to tighten up the firewall as much as possible, you could do your testing on a server that is more accessible, and once you've proven the configuration then just copy that configuration to your production server.
When you're working on a remote system and you're worried about making a configuration change that could lock you out, it's a good idea to have something out-of-band that lets you recover. Examples:
- A KVM over IP device so you can access the local console even with networking disabled
- HP iLO or Dell DRAC management technology
- Somebody you can call who has physical access to the server
If you don't have those available, you can usually set up some kind of failsafe. In your situation, try creating a scheduled task that runs netsh advfirewall reset. If you lock yourself out with a firewall rule, just wait until the scheduled task runs. Test your failsafe beforehand to ensure that you can rely on it.
