================================================================================================================================================================================
FROSTYVERSE TEST H2-091 — FROSTCELL NODE OWNERSHIP / 4D MULTIPLICITY ATTACK
================================================================================================================================================================================
version: 2.0.91-frostcell-eight-node-four-face-4D-ownership-attack
patch:   h2-090-eight-node-halfD-primary-four-crossed-face-fullD-sidequest-no-double-count-1
baseline H2-090 sha256: 374d28de285ee457e2764ae187a5331538e7172c524348fd5edc873e3d98ca6c
python:  3.10.6
profile: full samples= 200

TEST 01 - CONTRACT / FALSIFICATION ORDER
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
  R01 [PRIMARY-8-NODE] Preferentially test the user hypothesis: all eight corner FrostNodes of the deformed FrostCell each independently own one H2-090 source/destination half-ledger (1/2 D).
  R02 [FACE-SIDEQUEST] Separately test four crossed-face FrostNodes x one complete D; the crossed face must be selected by transit geometry, never by hand.
  R03 [TRILINEAR-GEOMETRY] Use the inherited eight-node deformed FrostCell map to establish geometric participation and boundary-node count.
  R04 [NO-DOUBLE-COUNT] Decompose the actual inherited H017 twelve-FrostSegment cell work, count every physical cell segment once, and split each segment symmetrically between its two endpoint nodes.
  R05 [SIX-SUPPORT-INVENTORY] Inventory all six incident supports around all eight corners, including external supports, but do not invent a new whole-cell activation law for the external 24 segments.
  R06 [DEFORMED-CELL] Repeat node-share conservation on inherited named deformed FrostCells; no flat-cube-only conclusion.
  R07 [FV-ONLY] No GR, Mercury target, 4D target amplitude, or pair-mass score selects the mechanical outcome.
  R08 [PAIR-RULE-FROZEN] H2-089 M1+M2 pair-total bookkeeping remains frozen and is not retested here.

TEST 02 - EIGHT-NODE DEFORMED-FROSTCELL GEOMETRY / FACE SELECTION
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
  deformed/flat cell cases                               = 6
  unique-face cases: all 8 nodes nonzero through interior= True
  unique-face exits: boundary nonzero nodes exactly 4    = True
  edge/corner controls assigned an arbitrary face         = False
  max trilinear partition-of-unity error                  = 2.220e-16

TEST 03 - SIX-SUPPORT NETWORK INVENTORY AROUND ALL EIGHT CORNERS
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
  8 corners x 6 incident support endpoints                = 48
  unique physical segments incident to those corners      = 36
    internal FrostCell edges                               = 12
    external neighboring segments                          = 24
  inherited H017 active FrostCell segment frame            = 12
  inherited H017 whole-cell activation law for external 24 = False
  structural endpoint transverse sum (all 8 six-supports)  = 32.000000000000
  unique-physical-segment transverse sum                   = 24.000000000000

TEST 04 - PRIMARY 8x(1/2D) + FOUR-FACE x D NO-DOUBLE-COUNTING WORK AUDIT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
  actual all-8 source-half shares / D                 = 0.500000000000 .. 0.500000000000 mean=0.500000000000
  actual all-8 destination-half shares / D            = 0.500000000000 .. 0.500000000000 mean=0.500000000000
  actual all-8 full handoff shares / D                = 1.000000000000 .. 1.000000000000 mean=1.000000000000
  replicated candidate 8 x (1/2D) / D                 = 4.000000000000 .. 4.000000000000 mean=4.000000000000
  actual mechanically-selected face-four full shares / D = 0.500000000000 .. 0.500000000000 mean=0.500000000000
  replicated SideQuest candidate 4 x D / D                = 4.000000000000 .. 4.000000000000 mean=4.000000000000
  NOTE: replicated candidates count only if they match the independent 12-segment cell ledger.

TEST 05 - DEFORMED FROSTCELL NODE-SHARE CONSERVATION STRESS
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
  deformed/flat work cases                                 = 36
  max |sum of 8 endpoint node shares - 1|                  = 2.220e-16
  node-share range across deformed cases                    = 0.055418 .. 0.245529
  crossed-face four share range across deformed cases       = 0.478494 .. 0.518922

TEST 06 - H2-091 DECISION
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
  primary outcome: B_ALL_EIGHT_NODES_GEOMETRICALLY_ENGAGED_BUT_CURRENT_CELL_WORK_PARTITIONS_ONE_D
  all 8 FrostCell corner nodes geometrically engaged in interior: True
  actual no-double-count all-8 full handoff / D = 1.000000000000
  arithmetic 8 x (1/2D) candidate / D         = 4.000000000000
  candidate / actual current cell ledger       = 4.000000000000
  SideQuest outcome: B_CROSSED_FACE_FOUR_SELECTED_BUT_CURRENT_CELL_WORK_IS_PARTITION_NOT_4D
  crossed-face four actual full-work share / D = 0.500000 .. 0.500000
  edge/corner exits avoid arbitrary face selection: True
  external support whole-cell work law currently missing: True
  interpretation: the user intuition that all eight nodes belong to and geometrically define the deformed FrostCell is SUPPORTED.
                  However, in the CURRENT inherited H017 cell-work ledger those eight nodes partition one D-scale complete handoff; cloning the H2-090 half-ledger onto every corner counts the same cell work multiple times.
                  The eight-node 4D idea would require an additional derived activation/ownership law for the surrounding external node supports; H2-091 does not invent one.
  H2-089 pair-total mass rule changed by H2-091: NO
  SEALED H2-091 FV-only fingerprint: e2a0728d22f39965ae420bf82f2d1dae18b116e5d7296a9550c07e64325bc4da

H2-091 complete. Return h2-091results.txt first; CSVs are useful for node-ownership verification.
