nx_tissuematerial_gate.nx

nx_tissuematerial_gate.nx -- THE LAYER STIFFNESS RATIO IS CITED, AND THE NUMBER IT REPLACED IS PROVEN TO HAVE BEEN OUT OF RANGE. (softbody SB4, first leg, 2026-08-25.) WHAT THIS EXISTS TO STOP. nx_softtissue's per-layer distance compliance used to be two hand-tuned constants carrying the names ST_MAGIC_1229 and ST_MAGIC_12288 -- renamed by nx_magic apply, never sourced. A MAGIC NUMBER GIVEN A NAME IS STILL A MAGIC NUMBER: the grep goes quiet and the arbitrariness is untouched, which is strictly worse than a bare literal because it now LOOKS deliberate. Their implied gland/adipose stiffness ratio was 9.998 against a published within-study band of 1.420..5.698 -- 1.75x above the top of the literature, and nothing in the estate could see it. THE BAND IS READ FROM knowledge/gamefeel_oracle.conf, NEVER COMPILED IN. If the row is missing this gate REFUSES rather than passing on a default: a material gate that cannot find its citation and acquits anyway is the exact defect it was built to catch. WHY A RATIO. Elastography moduli for one tissue span ~50x across studies (the source review's own table carries adipose at 0.33, 3.3 and 17.1 kPa). That spread is METHOD, not biology, so absolute values are not transferable and none is banked. The within-study ratio survives because the method cancels on both sides of the division. license_tier: ORIGINAL No hw writes (Rule 26). expect_exit: 0