There are no energy sources. It's all just conversion from one form of energy into work.
That's kind of what I meant. How can you convert plasma (massive amount s of heat energy) into work (something usable)? What I meant with the steam engine thing is that even in nuclear power plants, all we really do is use the heat to boil water, so it's still just a glorified steam engine.
The thing is that fusion reactions creates enormous amounts of heat and some designs use plasma, so is there a better way to turn all that heat energy into work than boiling water?
There are ways.
You can use it to heat another substance, which can then be circulated in a heat exchanger that can be found in pretty much all electric power plants based on producing heat, be it with burning, nuclear power, concentrating solar radiation with mirrors etc. Of course this is the glorified steam engine you speak of, but what isn't then. It's a good way, but with fusion reactor the temperature tends to be a problem of some extent, more so than on fission based nuclear reactors.
To deal with extreme temperature, which normal matter obviously wouldn't hold, fusion reactors that are known to work for some time use a torus-shaped magnetic field to contain the plasma where the fusion actually happens, to separate it from physical contact with the reactor walls. Of course the radiation will heat the inner walls, but that can be dealt with, if enough energy is transferred away from the core. Heat exchanger does exactly that - it cools the reactor core and heats up itself, it's pressure increases and you got yourself thermal energy from the fusion.
If you want to avoid the steam engine altogether, another possibility in case of fusion is that you can use induction to capture the kinetic energy of the charged particles coming from the fusion (mainly alpha particles, aka. Helium.4 nuclei). After all, movement of charged particles is current, right? You just need to make sure that the particles are guided to same direction via static magnetic fields, and all is fine and dandy - they will cause a current to induce to nearby coils.
Partially, you could also possibly use some kind of layer of photoelectric cells to try and capture the gamma rays from the reaction, but the problem with that is that you need a really thick stack of the cells due to high penetration of gamma rays.
Of course, if at some point the integrity problems are solved, you could basically build a frakking four-stroke internal fusion engine that replaces spark plugs with, say, a laser and the fuel injection system would spray a small cloud of plasma into a chamber, where it would be contained in the middle of the chamber... Then, bang the laser heats the plasma sufficiently to start a fusion reaction, the plasma heats up a lot, expands, pushes the cylinder down. Exhaust would consist of helium; the "only" problem would be the stray neutrons making the engine block and other components radioactive on slow or not-so-slow rate, depending on what kind of reaction was used.
Of course, seeing how four-stroke engines are most definitely not amongst the most thermally efficient devices, if there actually were a need to convert energy from fusion directly to kinetic energy, a stirling engine would be more viable solution.
By the way, it's a common misconception that plasma must be really hot and fiery and melt everything it touches instantly. It doesn't really do this because in most applications, plasma is not really that dense. But to achieve fusion, the plasma needs to be heated to really high temperatures to offset the lack of pressure - you can't very well compress 15 000 000 Celcius-degree plasma mechanically. Magnetic compression is a possibility, but it takes a lot of energy in itself. Thus as far as I know, the fusion reactors currently designed will use greater temperature than in, say, Sun's core, to replace the missing pressure element that is present on natural fusion reactors, stars, where the massive gravity well makes the static pressure pretty much uncomprehensible on the core.
