nx_ecomat_evprobe.nx

nx_ecomat_evprobe.nx -- DOES EACH DOMAIN'S EVIDENCE POINTER ACTUALLY MATCH ITS LOG? ★WHY (measured 2026-08-07). A domain's grade is live-derived by scanning its evidence log for its declared pattern. em_derive_level matches those bytes EXACTLY. When the pattern's CASE is wrong the scan finds nothing and the derive returns TOY(1) -- and TOY is indistinguishable, in every report downstream, from a real measurement of a genuinely immature domain. CAUGHT LIVE: lab-science declared "verdict=" while knowledge/status/labsci_gate.log emits "VERDICT=GREEN checks=13 passed=13 permil=1000". A domain scoring 13/13 at permil=1000 was being published as TOY, and the ecosystem headline carried that number. ★A CONSTANT WEARING THE SHAPE OF A MEASUREMENT IS WORSE THAN A BLANK: a blank would have shown up as DANGLING and been chased. This showed up as a grade. So: for EVERY domain and BOTH evidence slots, report whether the declared pattern is actually PRESENT in the declared log -- and when it is absent, say whether a CASE-VARIANT of it is present, which is the difference between "this gate has not run" and "your pattern is misspelled". Full population, never a sample: every domain, every declared slot, counts printed. nx_ecomat_evprobe [store_prefix] exit: 0 all declared pointers MATCH | 1 at least one ABSENT | 2 store empty/missing license_tier: ORIGINAL expect_exit: 0