I really didn't want to get into this but it's getting to be too much.
First off, everything I am going to talk about is not the specificly designed race application parts. Short of that, stuff that was designed and built for a race team, almost everything out there for commercially available automotive applications are improperly designed and badly executed. There are exceptions, Seibon spends a ton of money making parts that work as designed, the price reflects that. Companies making nock-offs just take their design and steal it.
Properly designed carbon parts are NOT brittle. The carbon fibers themselves have an infinite strain to failure. This means that as long as you do not strain it past it's yield point it will never break, deform, or change it's properties. In plain english, if you put a steel bar under tension (pulling) and apply and release a force (that doesn't break it or cause it to pull into two pieces, yield) and cycle it thru this loading and unloading, it will eventually strech and break over a long period of time. Steel has a very low strain to failure rate but a high yield point, it's able to bend and contort without breaking. It can also carry a very high load without breaking initially but will eventually change shape. Carbon fiber is basicaly the opposite. As long as you only apply (or load and unload) a force that is less than what it takes to break it, it will strech and return to its orignial shape forever. But there is a problem with this.
The five fundamental forces are:
Tension
Torsion
Bending
Compression
Shear
While steel has good properties for all five forces, with bending probably being the weakest (fyi, concrete is great in compression but very bad in torsion, bending, and shear), carbon fiber is really only useful when it is under tension. Think of trying to bend a single carbon fiber, it won't support much load. So when any other force but tension is applied to a carbon fiber part, one that also includes resin, it is actually the resin that is carrying the load. Under compression, the carbon isn't doing anything, it's the resin. It's basicly the same for all the forces but tension.
So, back to being brittle. The example of a skid plate is actually a very good one. What is really needed is high shock resistance, we have a two ton car hitting an "immovable" object (a rock) at 35,60,90 mph. A metalic (steel or aluminum) plate is going to yield, bend and distort, but probably not break. It will probably keep you oil pan from getting a big old gouge or dent in it by getting torn up itself. It's a sacrifical item before any really important parts.
If we have that same impact hitting a carbon fiber plate, what is actually happening is that the resin is alowwing the structure (plate) to deform until the carbon fibers are placed under tension and are able to carry the loads (up to their breaking point). As long as the carbons yield strength has not been exceeded and the fibers have not been scratched or otherwise compromised, the structure maintains it's effectiveness and will work properly in the next impact. However the resin will be broken and cracked. If the yield strength is exceeded, the fibers will break and the structure is now useless.
The first problem is determining what kind of forces to design your structure, or plate, to withstand. The impact of a car hitting a bolder at 25mph is VERY different than 75mph. The plate for the 75mph crash will need to be signifigantly thicker than the 25mph. (I am not positive but I don't think that simply tripling the thickness would be sufficient). It may be that, for this example of a skid plate, the carbon will end up being heavier than an aluminum plate would need to be. Remember that if it isn't think enough it will be useless and your car will still get banged up.
The next problem is designing a structure that will not deform too much and end up transfering the load to your oil pan. The thicker, and heavier plate (than the aluminum) may be able to resist the impact but may bend too much.
Something that is brittle has a very low strain to failure rate, actually most resins are very brittle. They have a low tensile strength but are better in compression, they are also very rigid. This is why they are paired with something that is so good in tension but bad in compression, carbon. Most people think of carbon fiber being brittle because they have only ever seen the pieces that are for decoration, such as a pulley cover. They are very thin, have very few layers of carbon and probably have a very cheap resin. If you take a panel like that and bend it, of course it is going to crack the resin, but you probably won't be able to break the carbon fibers without some significant effort. That behavior is what carbon\resin composites are designed to do. A properly designed skid plate could withstand you taking a sledgehammer to it all day and you would probably only screw up the surface resin and the first layer of carbon. (fyi, if an individual fiber is damaged or compromised in any way, that fiber is no longer able to bear the same load, in the same way, as an undamaged fiber. This is why carbon is usually manufactured as a net shape and not machined)
Real world applications, a pulley cover is not load bearing and can be very thin, and needs to be in order to be light. A hood is also not load bearing and can be only a few layers of carbon, less than 1/16" and do it's job (which is covering the engine, nothing more). But all the hoods that I have seen are not made that way. They are the OEM shape, thickness and have the same structure on the underside that the metal ones need for support. Add in the fact that, most likely, the cheap ones are nothing more than one layer of carbon on top for looks and the rest is heavier fiberglass mat and I am not the least bit surprised that they are heavier than OEM. Real racing hood are probably exactly that, just a couple layers of carbon held on by hood pins. I have an old school fiberglass hood for my datson and while I haven't weighed both of them yet, I believe the glass one is going to be heavier. Same thing for my fenders.
Designing a carbon fiber part to meet all its needs is not cheap and if there is little or no design work done, the parts just aren't being made right. My opinion is that 99.9% of stuff out there is useless and a waste of money because of this fact.
It's late, I hope more coherent than most of Boxer3Main's posts.