The typical protein crystal does not withstand mechanical stress equally well than e.g., the ones of table salt, or refined sucrose. Hence the varying shapes of grains.
Photos about nice protein crystals by shape (habit) & extinction (check if you can use linearly polarized light / "crossed Nicols") often are recorded while they still are in their mother liquor (to grow them e.g., by sitting drop, hanging drop, or in capillary [see e.g. here] to mention some techniques) typically grown for in-house/beamline X-ray diffraction analysis. Here, you want them to be representative of your sample (e.g., in chemical composition, and crystalline phase), and yet good enough for subsequent data collection and data processing.
The commercial (perhaps microcrystalline) powder you have access to, is prepared with the primary intent to offer you access to a sample chemically (by molecular composition) pure enough. That is, concentrations of remaining solvent of reaction / purification, unwanted ions of heavy metals are below an (often arbitrary) threshold set. If there once were crystals, they now are broken in a more or less random way while passing sieves, being scrapped-off from filters, dried (think about solvent molecules like water which have to leave the solid), filled into a bottle which upon transport is shaken / crystals bounced against each other. This reduces your chances to meet a crystal with prominent cleavage planes. Instead, you find rough and conchoidal surfaces. In addition, the size of the grains may follow a more or less broad grain distribution, too.