RFM accumulator_copy_lens_binds_no_carrier: fixed by #13337, bisected to #12506; 4 residual reds re-measured

dag/gunbc/recurring_failure_mode/accumulator_copy_lens_binds_no_carrier.dag

@@ -7,9 +7,10 @@ data accumulator_copy_lens_binds_no_carrier: RecurringFailureMode = RecurringFai
   identity: "accumulator_copy_lens_binds_no_carrier" as NonEmptyStr,
 
   receipts: [
-    "**the accumulator-copy lens no longer classifies planted inline-fold copies: it reports no Poly2Suspect where one is owed** (INVALID STATE: a supported positive loses its discriminating classification). `v2.test.long.accumulator_copy_fold_analysis` runs 5 of 24 on main (claim_batch, BuildBuddy, 2026-10-04, swift-bat-688). Passing: `quadratic_ingest_has_lowered_loop`, `quadratic_ingest_has_no_surface_fold_call`, `identity_fold_has_no_finding`, `out_of_iteration_append_is_out_of_domain`, `item_inside_computed_argument_stays_clean`. Failing, 19, a MIXED population: positives (`quadratic_fold_has_suspect`, `real_branch_body_quadratic_is_caught`, `let_alias_is_proven_suspect`, `let_transitive_alias_is_proven_suspect`, the position specimens), refusal claims (`unregistered_combiner_reached_by_carrier_refuses`, `let_unknown_alias_still_refuses_may_alias` and others), at least one clean control (`let_fresh_binding_is_provably_clean`), and `report_counters_state_the_domain`, a conjunction over two snippets. One mechanism for all 19 is NOT established.",
-    "WHAT A FAILED HAS-SUSPECT DOES AND DOES NOT SHOW. `v2.lens.complexity_accumulator_copy.analyze` `analyze_lowered_loop`, and the surface-fold branch of `analyze`, answer an absent carrier with `Unclassifiable { cause: ^fold_accumulator_unread }`; `finding_is_suspect` is false for it and `finding_is_refusal` true. So the lens most likely REFUSES rather than answering clean. This is a loss of supported classification, loud at the lens. Whether any downstream consumer discards that refusal is not established, and no such defect is assigned to the lens here. That the carrier is unbound is the leading reading, not measured per claim. Candidate cause, NOT bisected: gunbc#13028 (lambda-binder projection base as lexical reference).",
-    "RUNG FOUND AT: mitigatable, a typed refusal standing where a classification was owed; invisible because the module is outside the required gate and rostered in `gunbc.witness.witness_deferral_freeze`. NEXT TRIGGER, the capability: the lens binds the fold carrier and classifies Poly2Suspect for every supported inline-fold positive of this module, while its clean controls stay clean and its unsupported/refusal controls still refuse. Refusing the Loop by name does NOT satisfy it; the current absent-carrier branch already does that.",
+    "**the accumulator-copy lens found the lowered fold Loop but bound a carrier no body reads, so every planted inline-fold copy went clean** (INVALID STATE: a supported positive loses its discriminating classification). `v2.test.long.accumulator_copy_fold_analysis` ran 5 of 24 on main (claim_batch, BuildBuddy, 2026-10-04, swift-bat-688).",
+    "ROOT CAUSE, BISECTED (2026-10-05, quick-koi-242). Not gunbc#13028, the earlier candidate. The carrier half came from gunbc#12506 (861c829ac6, 2026-10-03), which minted the fold carrier as a generated slot (`v2.std.anonymous_binder` `fresh_fold_carrier`). `v2.lens.complexity_accumulator_copy.analyze` `lowered_loop_carrier_optional` still read the Loop's ^loop_carrier_edge, which now named that slot instead of the authored formal. The earliest unjustified boundary was the lens reading an encoding edge where it wanted a role. The port half came from gunbc#12382: a named actual became an Authored edge, and the lens read positional children only.",
+    "FIXED by gunbc#13337 (aae7da34ca, 2026-10-05). The carrier is now read by role through `v2.compiler.fold_lowering` `fold_loop_step` / `fold_step_carrier_formal`, the route infer types it by. Actuals bind through `v2.std.arrow_signature` `application_binding_plan`. The module grew to 28 claims in the same change, and measured on main at d7ac84233a (claim_batch, BuildBuddy, 2026-10-05) it runs 24 of 28. The 4 reds are NOT this class. Each is an expecting-red probe at `gunbc.explicit_witness_admission` with a capability dissolution, and each was re-measured as failing for its stated reason: `let_alias_is_proven_suspect` (ingest returns Absent on the specimen; the grammar's let-then-call shape), and `named_step_fold_refuses_accumulator_unread`, `report_counters_state_the_domain`, `declared_named_step_fold_copy_is_caught` (the named-step specimen ingests and count_fold_sites returns 0; classes `fold_shape_visible_to_no_lens` and `call_expression_erased_at_v2_body_lowering`).",
+    "STANDING, in three parts kept apart. (a) THE HISTORICAL DEFECT WAS SILENT WRONGNESS, NOT A RUNG. For the supported planted copies, the lens returned clean where a Poly2Suspect was owed: a false negative with no containment mechanism. DESIGN section 4b puts that outside the ladder, not at mitigatable, and nothing executing caught it, because the module is outside the required gate. (b) THE REPAIRED BEHAVIOUR IS MEASURED, NOT GATED. The discriminators that go red if the carrier is read off ^loop_carrier_edge again (`quadratic_fold_has_suspect`, `real_branch_body_quadratic_is_caught` and the other inline-copy claims) passed under claim_batch at d7ac84233a. That is author-run evidence; no lane executes it. (c) THE EXECUTION OBLIGATION, SCOPED BY POPULATION. The declared drop `gunbc.rung_drop` `accumulator_copy_positive_controls_off_every_lane` covers ONLY its own six fold-shape controls, and its restoration does not establish coverage for anything else. `quadratic_fold_has_suspect` is enrolled as evidence on `gunbc.v2_compile_obligation_census` `accumulator_copy_root_lens`, and that row's trigger includes retiring the `test/claim/long/` witness-exclusion frontier. The module's other inline-copy discriminators, `real_branch_body_quadratic_is_caught` among them, are covered only by their `gunbc.witness.witness_deferral_freeze` `frozen_path_deferrals` row. That row is a debt marker that records identity only and names no executing cadence, so it is a record of missing coverage, not coverage. Missing capability: an executing lane (required or scheduled, with a declared ceiling) that plans the whole `v2.test.long.accumulator_copy_fold_analysis` module and reaches a verdict on every claim. That capability, not either cited row's retirement alone, is this class's NEXT TRIGGER.",
   ],
 
   evidence: [],