[K3] Cut the active-slice bridge at direct carrier inference

[scottdhughes]

Stack and review boundary

This is a ready zero-payment K3 semantic route-cut packet stacked on ready PR #1157 at exact head d7f0fd9370b3c13ff93293f08e03cadddb59b921.

Review only:

d7f0fd9370b3c13ff93293f08e03cadddb59b921..5145fc8e0abca6325b8226294cdc2661e0891dcc

Upstream main remains 93fba1be3f3299b0ba4708d88715377bbb656e45; #1157 is open, ready, and mergeable. After #1157 integrates, replay only this commit onto the integration head.

Result

This does not close K3. It identifies the first missing source theorem before the thirteen-route workboard and cuts one tempting shortcut exactly.

The active v4 sources do not define an executable actual-record predicate for

Z_BC^(m_in=2,r_out=4)

or a same-record selector/reconstruction/projection from an actual received line and bad affine slope to the supplied endpoint records. The packet freezes the required Rec_2_4, coverage, exact fiber, chronology, and all-Z_BC complement obligations. Raw endpoint labels or empty raw systems therefore cannot yet be charged as distinct affine slopes.

Over the deployed base field and actual order-2^21 carrier, the packet constructs 30 active and six source complete fibers of h(T)=T(1-T). They satisfy the printed degree-two source-pencil and rational-deck identities, but the displayed involution tau(T)=1-T does not preserve the carrier: 1 in D while tau(1)=0 notin D.

This refutes only direct endpoint-coordinate/evaluation-carrier identification. It is not an actual MCA received-line counterexample. The packet explicitly records that T-1/2 conjugates tau to T -> -T, which does preserve the even-order carrier; arbitrary record-preserving conjugated folds remain open.

Tangent source-pin repair

The exact three-file repair from #1139 commit 8d43c6fa3a6ff04ea369ba7046fced6ae133b097 is imported byte-for-byte. It binds the current grande_finale.tex source and revalidates the existing tangent atom:

U_paid                  = 981104
joint remaining reserve = 274980728110413983

This is a source-pin refresh, not a new payment. No #1156 scalar-locator file is changed.

The current MCA ledger remains:

U_paid = 981104
U_Q    = null
U_BC   = null
U_new  = null
K3 ledger movement = 0

LIST's U_list-int/U_ext chronology is not imported.

Historical alternative lane

The old equality-wall lane is pinned to its exact owner and reduction certificates. It is not a proved cap 68 and is not coupled to the current partition:

• it counts source-map equivalence classes per transversal projective residue line, not affine slopes;
• the exact excluded interval is 3912 <= delta <= 118076;
• five degree-12--16 low-excess windows and general excess remain open;
• each occupied direction has image cap 1,894,736 slopes;
• a current-partition adapter is still required.

Reconciliation and overlap

The packet explicitly reconciles #1130, #1132, #1139, #1143, #1152, #1155, #1156, and #1157. None supplies the missing actual-slope predicate, same-line reconstruction, exact projection fibers, or all-Z_BC complement fence. A refreshed open-PR audit found no duplicate packet or newer competing bridge.

Validation

• primary Python normal and -O: PASS;
• complete carrier census: 2,097,152 points, 1,071 internal tau pairs, 36 selected fibers;
• 19 local and 2 public-DAG source bindings: PASS;
• 43 semantic plus 3 parser mutations: rejected;
• all 92 Boolean-to-integer promotions and additional unknown-key/type probes: rejected in independent review;
• Sage exact finite-field/polynomial replay: PASS;
• FLINT independent polynomial/gcd/hash replay: PASS;
• connected Wolfram exact replay: PASS;
• local wm.sh -file syntax is correct, but the final repeat returned Insufficient credits;
• tangent verifier normal/-O and both tamper modes: PASS;
• tangent Lean package: builds, with only standard propext/Quot.sound (the older axioms=[] documentation claim is explicitly corrected in this packet);
• exact [K3] Cut the source-bound compiler at the active carrier bridge #1157 predecessor Python/-O/Sage/FLINT replay: PASS;
• schema validation, Python compilation, and diff checks: PASS;
• independent mathematics review: GREEN for this narrow route cut, RED for K3 closure;
• independent certificate/custody review: GREEN.

The JSON schema is intentionally only an envelope; the exact Python verifier supplies nested type/key enforcement. Historical commit replay requires non-shallow history containing 702cd8e1 and 065f347a.

Scope

This PR is ready at its stated scope: a source-definition/direct-coordinate route cut with zero ledger movement. It does not construct an actual received line, MCA witness, active-slice member, K3 orientation, projection-fiber bound, K3 allocation/payment, KoalaBear row closure, universal smooth-domain theorem, or official endpoint change.

The maximal next theorem is the actual-slope Rec_2_4 selector/reconstruction with exact projection fibers, chronology preservation, and the all-Z_BC complement fence.