Post-Close Episodic Write versus Lawful No-Write in a Tested Isolated Fixture

This paper studies a bounded post-close episodic-write comparison in the Garmon research architecture. After an episode has been closed, two independently assembled instances are compared on an isolated disposable fixture: Branch A performs a post-close episodic write, while Branch B lawfully withholds that write. Five pre-specified same-episode projections are compared. The branches do not differ on those five projections in the published evidence. Separately, reachability is reported as a human classification of observed touchpoints. The public checker recomputes the A/B projection equalities and the presence/absence of the trace, but does not recompute reachability or execute Garmon. The corresponding B4 Evidence Package v1 is publicly available at: https://doi.org/10.5281/zenodo.22813870