Linear blend skinning
Deforms each vertex as a weighted sum of the influencing bones' matrices — the default method for skeletal animation, fast but collapsing under twist.
Linear blend skinning (also skeletal subspace deformation) is very nearly all the skeletal deformation you have ever seen. Each vertex carries weights toward a handful of bones; its final position is the sum of every bone's transform applied to the vertex, scaled by the weight. The arithmetic is trivial and drops straight into a vertex shader, which has made it the default since Magnenat-Thalmann and colleagues in the 1980s.
Its flaw is as famous as the method: when a joint twists hard the matrices blend component by component and the volume shrinks — the "candy-wrapper" effect at shoulders and wrists. That is precisely the problem dual quaternion skinning was invented to fix.