Investigate the cadkernel path for ACIS kinds; add spikes and findings

OCS already creates, transforms and combines bodies through cadkernel, and
solid_to_sat / kernel_acis_body verify each result lifts back losslessly.
Two ignored spike tests measure volumes and DWG/DXF round trips. The
blocker is an operation channel, not geometry; the note recommends an
additive solid_operation host method and phases the work.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

docs/plugin-host-model-coverage-ledger.md

@@ -44,10 +44,10 @@ Completion gate codes: **C** = create through Python document model; **R** = rea
 | ViewBorder | **Integration-tested / blocked at W by acadrust (DXF kind)** | Writable: `min`, `max`, `center` (2-D pairs), `scale`, `rotation_angle`, `active_viewport` and `scale_handle` (both required at creation). `version` is read-only. Reference rules: `active_viewport` must be an existing Viewport entity and `scale_handle` an existing `Scale` object. Host validation requires finite bounds with `min` below `max` on both axes, a positive scale and a finite rotation. `staged_python_view_border_lifecycle_over_real_ipc` (shared `run_mesh_lifecycle`, driver-made viewport and scale fixtures) passes C/R/E/D/U/I/V/P and DWG W: creation, a resize/recenter/scale/rotation transaction, canvas wires, seven rejection cases plus an atomic rollback, DWG reopen and re-edit, and three-step undo/redo. **W blocker (engine, pinned by a canary):** a DXF `DRAWINGVIEW` reopens as a different, opaque entity kind rather than as a ViewBorder, so only DWG restores the typed record. Not `Complete` until DXF loading restores the typed entity. Not covered: the linked view representation and drawing-view definition records. |
 | SectionSymbol | **Blocked** | Not mapped. A valid section mark needs a `SectionViewStyle` and a `ViewRep` view-representation object that a script cannot create through the host API, and the entity stores raw point counts and flag words (`raw_point_count_90`, `raw_flags_90`, `raw_point_record_count`, `raw_view_symbol_70`) whose relationship to `points` is undocumented in acadrust and is written back verbatim by both writers, so the host cannot keep them consistent after a scripted edit without guessing. Adding the kind needs (1) host-side creation or lookup of a section style and view representation, and (2) a confirmed derivation of the raw counts from a real DWG produced by another CAD application. The kind keeps raw snapshot access and layer edits. |
 | Light | **Integration-tested / blocked at W by acadrust (DXF `cast_shadows`)** | Writable: `name`, `light_type` (1 distant, 2 point, 3 spot), `position`, `target`, `status`, `light_color`, `plot_glyph`, `intensity`, the attenuation type and limits, `hotspot_angle`, `falloff_angle`, `cast_shadows`, `shadow_type`, `shadow_map_size`, `shadow_map_softness`, `photometric_mode`, `photometric_data`; `class_version` is read-only. Host validation requires a non-empty name, a known light type, finite placement, a spot or distant target that differs from the position, non-negative intensity, angles and limits, an attenuation start not above its end, a spot hotspot not above its falloff, and a non-negative shadow map size. `staged_python_light_lifecycle_over_real_ipc` (shared `run_mesh_lifecycle`) passes C/R/E/D/U/I/V/P and DWG W: a spot light with an RGB colour, an intensity/target/position/shadow transaction that preserves the colour and untouched fields, the canvas glyph, eight rejection cases plus an atomic rollback, DWG reopen and re-edit, and three-step undo/redo. **W blocker (engine, pinned by a canary):** the DXF reader does not restore `cast_shadows`, which reopens false; colour, placement and intensity persist. Not covered: photometric web data edits and sun/skylight objects. |
-| Region | **Blocked** | Not mapped. The entity is an opaque ACIS payload (`acis_data` SAT/SAB text or binary) plus derived `wires`, `silhouettes` and, for solids, a history record. A script can neither produce a valid body nor edit one, because every geometric change needs the CAD kernel to rewrite the payload, and the handover forbids deserializing and rewriting opaque ACIS data without a proven lossless path. The only lossless edit today is the layer change the document model already supports. Adding this kind needs a host-side, kernel-backed API (create a primitive, or apply a transform, then regenerate `acis_data` with its wires and silhouettes) with a DWG and DXF round-trip that preserves the payload and history byte-for-byte for untouched bodies. |
-| Body | **Blocked** | Not mapped. The entity is an opaque ACIS payload (`acis_data` SAT/SAB text or binary) plus derived `wires`, `silhouettes` and, for solids, a history record. A script can neither produce a valid body nor edit one, because every geometric change needs the CAD kernel to rewrite the payload, and the handover forbids deserializing and rewriting opaque ACIS data without a proven lossless path. The only lossless edit today is the layer change the document model already supports. Adding this kind needs a host-side, kernel-backed API (create a primitive, or apply a transform, then regenerate `acis_data` with its wires and silhouettes) with a DWG and DXF round-trip that preserves the payload and history byte-for-byte for untouched bodies. |
-| Solid3D | **Blocked** | Not mapped. The entity is an opaque ACIS payload (`acis_data` SAT/SAB text or binary) plus derived `wires`, `silhouettes` and, for solids, a history record. A script can neither produce a valid body nor edit one, because every geometric change needs the CAD kernel to rewrite the payload, and the handover forbids deserializing and rewriting opaque ACIS data without a proven lossless path. The only lossless edit today is the layer change the document model already supports. Adding this kind needs a host-side, kernel-backed API (create a primitive, or apply a transform, then regenerate `acis_data` with its wires and silhouettes) with a DWG and DXF round-trip that preserves the payload and history byte-for-byte for untouched bodies. |
-| Surface | **Blocked** | Not mapped. The entity is an opaque ACIS payload (`acis_data` SAT/SAB text or binary) plus derived `wires`, `silhouettes` and, for solids, a history record. A script can neither produce a valid body nor edit one, because every geometric change needs the CAD kernel to rewrite the payload, and the handover forbids deserializing and rewriting opaque ACIS data without a proven lossless path. The only lossless edit today is the layer change the document model already supports. Adding this kind needs a host-side, kernel-backed API (create a primitive, or apply a transform, then regenerate `acis_data` with its wires and silhouettes) with a DWG and DXF round-trip that preserves the payload and history byte-for-byte for untouched bodies. |
+| Region | **Blocked, unblocked by design** | Not mapped yet. Transform uses the same pipeline; creation needs a "region from these curve handles" operation built on the REGION command's `add_region_model`. See [cadkernel-body-path.md](cadkernel-body-path.md). |
+| Body | **Blocked, unblocked by design** | Not mapped yet. Same pipeline as Solid3D: `kernel_acis_body` lifts a `Body` payload and the spike round-tripped it through DWG and DXF. A genuinely non-solid (sheet or wire) body may not lift and would stay layer-only. See [cadkernel-body-path.md](cadkernel-body-path.md). |
+| Solid3D | **Blocked, unblocked by design** | Not mapped yet. The 2026-09-20 kernel investigation ([cadkernel-body-path.md](cadkernel-body-path.md)) shows OCS already creates the payload (nine primitives), transforms it and combines bodies through cadkernel, and that `solid_to_sat` and `kernel_acis_body` verify every result lifts back with no loss; spikes confirmed exact volumes and DWG and DXF round trips. What is missing is a script-to-host operation channel (recommended: an additive `solid_operation` host method covering create-primitive, transform and boolean), not geometry. Open items: boolean speed (about 23 s in a debug build, unmeasured in release), history-carrying solids, and AutoCAD acceptance, which cannot be tested here. |
+| Surface | **Blocked, unblocked by design** | Not mapped yet, and the most involved: a surface carries `surface_data` beside `acis_data`, and only a plane keeps its kind when regenerated. Plan: planes first, then extruded and revolved from a profile. See [cadkernel-body-path.md](cadkernel-body-path.md). |
 | Ole2Frame | **Update-only (all applicable gates pass)** | Writable: `upper_left_corner`, `lower_right_corner`, `lock_aspect`, `is_paper_space`. Read-only: `version`, `source_application`, `ole_object_type`, `dwg_mode`. The embedded object `storage` and the frame `envelope` are opaque, excluded from the Python form, refused by the mutation check (`Ole2Frame embedded storage cannot change`) and preserved untouched. Host validation requires finite corners spanning a nonzero width and height and `lock_aspect` of 0 or 1. Creation from a script is refused with an explanatory message, because a valid frame needs a real embedded object; the host's own `build_embedded_ole` builds one from an image. `audit_python_ole2frame_lifecycle_over_real_ipc` (shared driver; the fixture embeds a real 4x3 PNG through `add_embedded_image`) passes E/D/U/I/W/V/P: a corner and aspect-lock transaction, canvas geometry, rejection of a zero-size frame, a bad aspect lock, non-finite corners, a read-only `version` and an attempt to write `storage`, an atomic rollback, **DWG and DXF reopen with the 119-byte embedded picture byte-for-byte intact (extracted through `extract_presentation`)**, re-edit, and two-step undo/redo. Not `Complete` because Python cannot create a frame. A create path (`file_path` of an image, built with `build_embedded_ole`) would need a manual-kind converter like Dimension and is the natural follow-up. |
 
 After each increment, update the row with test names, fixture paths, exact