I suspect it's more that certain authors enjoy a much higher or lower powered game than the average, and this tends to show up in their work. In addition, there's the possibility that they simply didn't realize something would be as weak or strong as it ends up being- the truenamer couldn't have been a deliberate calculation, but maybe the planar shepard fit right into the group it was *ahem* playtested in.

It's best to think of even official splatbooks as simply collections of homebrew material from people who are paid to do it. You're going to get a much more thorough take on something than you'd usually find online, but it's usefulness and potential to break the game are things you just have to judge on an individual basis.