By the time I wrote this, twenty numbered attempts at one Phebs scale gate had been consumed. None had established the full claim. That does not mean I ran the same experiment twenty times. Some attempts stopped during plan review or preflight. Some ran for hours. Some produced a signed, source-free receipt; others found that the evidence pipeline itself could not make a trustworthy record, so they sealed nothing and retired the identifier. Take 18 completed the two-million-file structural...