The soil testing is a moot point. Whoever does the test, if they know what they're doing, will have to put togther a graph based on the soil density obtainable through a standard process for a range of soil moistures. The problem is, it sounds like the material was too wet when it was placed. That means the pore pressure was too high which might account for the sponginess the builder reported. I'm guessing that the backfill material will be contained so over time the second story garage will not be undermined when the clay settles or washes out from the bottom. I've ranted about using clay as backfill in many of the retaining wall posts. Make no mistake about it, It sounds like at least two walls of your walkout basement as currently planned will be retaining walls.
You'd be better off reworking the design to use a precast slab for the garage floor and leave the area underneath open. If you're using block for the walls, there are some other things the builder needs to do. It may cost more now, but if it isn't done right the problems will show up at some point. In some cases I've seen the cracks show up not long after the walls were laid. Over time, as happened to a friend, a basement wall can collapse which is a real PITA if you're using it as a finished family room and you end up with several tons of mud lying on top of your carpet not to mention the new view of the outdoors.
Rather than use block, if that's what you're planning, there are a lot of concrete contractors in MD and maybe closer that can construct a basement of reinforced concrete. In many areas the reinforced concrete is cheaper than block, especially if masons are few and far between. That will eliminate a lot of the potential problems. If you go that route make sure they install a waterstop in the basement floor slab to tie into the walls.
The other point to keep in mind is that the 100 and 500 year floods are mathematical constructs. They are statistical calculations based on the history of flooding in your area. Since statistics works best when you have a large number of samples, it's obvious that a hundred years of records isn't enough. Even in Europe with 300 or 40 years of records they get surprises. In my area consistent recordings go back to 1929. We've had three floods that exceeded the 100 year flood since then. The 85 flood was beyond anything they could calculate. Keep in mind a 500 year flood in most cases isn't much higher than a 100 year flood. The recent flood in NJ east of the Cherry Hill area resulted from 11" of rain. One newspaper account termed it a 1000 year flood. That was BS. I doubt given the circumstances, several dams failed, someone actually did the math.
In your case, I'd head for the local NRCS office which should be located in or near the county seat. they have a free booklet on designing ponds that goes into how to calculate runoff. The local Farm Service Agency which may be co-located with the NRCS office will have aerial photos of your area. They also have a device called a planimeter that they can use to calculate the acreage upstream from you. It wouldn't be hard to run some worse case calculations and see just what might come downstream. The booklet gives you all the equations and examples to follow. I'd plug in at least a 10" rain. With maybe an hours work, you'll have something to think about.
This may seem like a lot of effort, but there're folks in many areas of this state that have recently seen flash floods that have exceeded anything before. BTW, since you're not in a flood zone, you can buy flood insurance at the preferred rate which is around $100 per year. You might want to do that based on the results of your calculations.
Another factor to consider is unless you own all the upstream property you can't guarantee what may come downstream one day. Whether it's brush or childrens' toys. Your culvert might end up being a dam once it's blocked. Where does the water go then? Water can undercut a foundation very quickly that close to the creek.