Not related to LEDs nor the library, but for those of you wanting to venture into creating your own boards for driving LEDs, this is what bit me these past few days. It wasn’t till I brought the model in 3D space that I noticed the pads were reversed. Pads 1 and 3 are the oscillator, and grounding them will, well, simply not work. 
Chalk this one up to copying from another library and making the assumption it was correct to begin with. All is not lost though. Simply rotating the crystal 90 degrees will fix this issue on the test boards. The new ones will have the pads oriented correctly. So I can still use these test boards.
Haha, I ran into this same eagle lib bug too.
It helps when you create accurate models too. The old crystal model I had did not have the key indicator and was the wrong size, so I had to create a new one. By having the key on the pad, I discovered the problem. Imagine if they were all the same, no key.