From 560c2d0824bd4974188dc242dea0ae4db7aea230 Mon Sep 17 00:00:00 2001
From: Brummel <mould@mycelium.de>
Date: Sat, 13 Jun 2026 19:08:54 +0200
Subject: [PATCH] =?UTF-8?q?audit:=20cycle=200035=20=E2=80=94=20drift-clean?=
 =?UTF-8?q?=20(node=20instance=20name=20in=20the=20graph=20model)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Close-audit for cycle 0035 (commit range 20000dd..0ae8320). Architect drift
review against the design ledger + CLAUDE.md: clean. No regression scripts are
configured (the architect is the gate); build + test + clippy verified green.

cycle 0035 tidy (clean):
- C23 preserved: the instance name is resolved to Option at authoring
  (instance_name()), emitted only on the render/model surface, and dropped at
  lowering — the compilat wiring is byte-unchanged.
- C14 preserved: model bytes stay deterministic; both golden twins (inline
  model_golden + sample-model.json) moved in lockstep to the same two "name"
  additions.
- C12/C19 untouched: the cycle adds no param-space surface; the param-ganging
  alternative was deflected to idea #61 (not via name collision — param_space
  is injective by design).
- Engine<->viewer lockstep intact: the conditional "name" producer (prim_record)
  and its three consumers (adaptNodes, genDot leaf-emit, cellLabel) moved
  together, protected by a prim_record unit test + a headless render guard.

The blueprint-viewer `SMA[length]` (a051571) and the compiled `Node::label`
`SMA(2)` are legitimately distinct surfaces (name-bearing blueprint model vs
value-bearing compiled view) — the split the ledger already draws at
INDEX.md:681-684, not drift.