A transposing instrument's WRITTEN key signature must be transposed along with its notes

When an instrument transposes, its part is conventionally notated with a key signature appropriate to its OWN written pitch (transposed the same interval as the notes), not the concert-pitch key signature — a Bb clarinet playing a concert-C-major passage is written in D major (up a whole step), key signature and all, not written in C major with clarinet notes that happen to sound a step low. This keeps every note's relationship to its own staff position, and to the accidentals implied by its key signature, internally consistent for the player reading only their own part.

A transposing instrument's written part carrying the CONCERT-pitch key signature instead of its own transposed key signature — the player would then need to mentally apply the transposition interval to every accidental implied by the key signature in addition to every written note, which defeats the entire purpose of a transposed part.

This is universal, uncontested orchestration/notation convention (standard in Adler, The Study of Orchestration, and implicit throughout Gould, Behind Bars' transposition discussion, though Gould's own chapter focuses more on which instruments transpose by which interval than on restating this base convention, which she treats as too basic to need restating). Ross and Read do not add anything beyond confirming it.

This rule can be decided from pitches, durations and voices — no layout required. A validator can check it before anything is drawn.