fully automated set

You are not providing enough information to get a clear idea about what you are asking. For example, how far apart is the Sky-Shed from your office?

There are a lot of limitations to what you can do because the CGEM is not designed for 100% remote operations.

You are going to need to do this incrementally making a decision, and then looking at the next level of equipment and software. A good place to start is what operating system and computer hardware platform are you using?

Assuming that between your two locations you can get a good WiFi signal, and you have line power at the observatory, then I would recommend a dedicated laptop PC in the observatory running 64b MS-Windows 7, and a Windows PC in the office with version 7 or later.  Set up the observatory PC as a Windows Remote Desktop server, and the office PC as a client.

WiFi because running network cable outside is too sensitive to electromagnetic interference and pulse surge. Windows 7 because you can run ASCOM drivers and that makes interfacing between hardware and software much easier; and later versions of Windows are not support by all the different vendors of hardware and software. A laptop with >=4GB RAM, >=250GB disk, >=2.0 GHz processor, dual core or more, and USB 3.0. A laptop is small, made for harsher operating environments, and fails to batteries when you have a power failure. 64b OS gives you access to all the memory. USB 3.0 allows some cameras faster downloads.

Most mounts, including the CGEM, have no absolute position sensing. At a minimum, a fully unattended system must have fail safe sensing of home, and RA and DEC limits. The limit controls must be able to shutdown the mount even if the hand controller, and any other control system, has either crashed or issued a problem command. Procedures you need to think through for CGEM remote operation:

1. CGEM has no absolute position sensors. It will loose it's GoTo calibration if it is not sent to home before powering off. When the system is powered on you need to check to see if the GoTo calibration is still valid by going to a star that is not to close to the mount's limits of travel. The worry is if the calibration is way off, then the scope might hit the pier. The best thing is to be at the scope at power up. Another alternative is to have a video camera in the observatory. Some people build their own limit switches.

2. Related to 1 is the need to have a GoTo calibration method. When this needs to be done, how quickly do you want to do it? If you want to do it quickly, then you may want a SkySync GPS to fetch the time, and a StarSense to be semiautomatic about calibration. Fully automatic calibration of a CGEM is not possible because there is no home sensor. Both devices tend to need a lot of sky, so that impacts how you are controlling the dome's open position.

3. The mount has no internal meridian flip routine. Some acquisition control software packages can automate this. Each has it strengths and limitations. This operation requires coordinating dome, mount, guide system, focus system, and camera. Here are three packages that might be of interest to you are:
Cyanogen Maxim DL
Software Bisque TheSkyX Professional Edition
Main Sequence Software  Sequence Generator Pro

There are as many remote control solutions as there are people because everyone has specific needs and different equipment. For example, most of my remote control needs are controlling a telescope on the balcony of my condo from my study inside when it is cold. However, most of my friends who are doing remote control have their telescopes at a remote site 50 mile outside of the city. When there is a problem, I run down stairs. When they have a problem, they hop in their cars.

Yes, hibernation is another name for parking before shutting down. Especially as you are learning you will find you will need to reset (power down everything) and realign when you make a mistake.

The only way to hibernate remotely is by using a utility called NexRemote. You must start up using NexRemote, disconnect NexRemote after mount power up, and then connect your other programs. At shutdown, you must manually disconnect all other programs, and connect to NexRemote to hibernate. Here is a link for a discussion about using NexRemote for hibernation.
You will find that NexRemote can be a challenge to install on later versions of Windows. This is one of the reasons I recommended a local Windows 7 PC to be remote controlled.

The sky model in the hand set is not perfect. At best, every few weeks, you will need to reset and realign. You can build a more accurate sky model on the PC with TheSkyX tpoint. I would not recommend starting with sky modeling as a beginner.

Everything is made harder the longer the focal length. Using a 8" at prime f/10 or with a f/6.3 or f/7 reducer takes experience because you will have too much magnification. A HyperStar will get an SCT short enough, but it introduces a lot of other things to learn about, and limits the choices of cameras. Most people are best off with a fast (=<f/7.0), short (=<500mm focal length), small (<=81mm) refractor to start.

Do not confuse polar alignment with GoTo alignment (aka calibration). Polar alignment makes sure the RA axis of rotation is parallel to the Earth's axis of rotation. GoTo alignment tells the computer the orientation of the mount and telescope relative to the heavens.

You will be able to send the scope/mount to park, save the settings by hibernating, and power down. On power up, if the scope has not been moved, you will be able to accept the saved settings, and begin. Problems take place under three circumstances:
1) The scope has been moved.
2) There was a power failure, or some other kind of tracking failure, before parking and hibernating.
3) After several hibernation cycles, or a few weeks even if not used, the mount's GoTo must be realigned.

You can set a software limit in the handset to have the mount stop tracking, and if you are guiding using the ST-4 outlet, it will also stop (this is a Celestron feature, not all mounts have software limits or monitor the ST-4 port). When the scope hits the limit, it will loose its alignment, and you have to realign the GoTo. The problem with the handset limit is it assumes you know what you are doing and you can see it. If you manually slew the scope locally, remotely through NexRemote, or remotely through the ASCOM driver, the scope will go pass the limit, and if the scope can hit something it will.

German equatorial mounts (GEM) cannot go much past the zenith without hitting something. Once you have gone past zenith one of three things must happen:
1) you did not set the RA travel limit correctly, and the camera hits the pier usually breaking something.
2) you set the RA travel limit correctly, and the scope stops, but looses its GoTo alignment.
3) you perform a meridian flip.

A meridian flip is performed in the following sequence either manually, or if the software supports it under software control:
1) stop camera from exposing
2) stop guiding
3) perform a go to the same position you were pointing, but on the other side of the mount
4) make sure you are pointing at what you are shooting, a stock CGEM, using Celestron's all sky polar alignment (ASPA) routine in the handset, and a standard 3+4 star alignment typically will not be accurate enough for centering a 8" f/10 scope with an APS-C sensor, so something must make a correction.
5) typically an SCT will encounter a focus shift called a "mirror flop" when a meridian flip takes place, and will need to be refocused (not all automatic focus systems will handle mirror flop, also some auto focus systems require pointing to a focus star on the same side of the mount as the target to refocus)
6) find a guide star
7) start guiding
start camera exposure sequence

There are several ways to point accurately to center the target:
1) make trail exposures and move the scope to the right place manually
2) use plate solving software
3) perform a highly accurate polar align no more than +/-2" error (very difficult, and requires drift alignment) and use software sky modeling. Sky modeling requires plate solving for its calibration, this is not an either/or situation.

Plate solving has the computer determine where the scope is pointing. Plate solving software needs at least 20 GB of disk space for star tables and a fast computer because it is computationally expensive. The sequence for plate solving is:
1) tell software RA and DEC coordinates desired.
2) software takes over imaging camera and takes a picture
3) picture is processed
4) picture is compared to a database of stars. This can take a long time with a slow computer.
5) software determines the relative difference between target and actual pointing
6) software commands mount to move to new relative position
7) until the accepted level of pointing error is reached the software will repeat steps 3 through 6
when error tolerance is reached software tells mount where it is pointing (this is called syncing the mount)
9) software notifies user or calling program

I have taking the time because hitting the tripod or pier with your camera or diagonal and eyepiece is a common unhappy experience made much worse when unattended. As you can see there is a lot of detail to learn. This is not the place to learn it. So I will leave you to your hands on experience.

can I get an opinion on SKX pro….or CCD Commander
with auto pilot ..