I don’t understand how exactly works the skew transform of a <node>. Please someone can explain me that and how can I compute this skew transform on a matrix ?

Second bonus question The transformations, need to be processed into theirs arrival order right ? But it is possible to find some transforms, next an instance of geometry or camera for exemple, next again some transforms and again an instance ?

about the skew implementation: it’s the same as the skew - transform found in the renderman specification. Ask google about that, and you might also take a look into this code: