Attach PPI pixels to documents

docs/modernization/build-inventory.md

@@ -92,7 +92,7 @@ Known local toolchain state:
   `pp_ui_core`, `pano_cli`, and their current headless test binaries,
   including foundation event/logging/task queue coverage, PNG metadata and
   decode, PPI header/layout, settings document, document
-  snapshot/per-layer-frame/move/duration coverage, paint
+  snapshot/per-layer-frame/move/duration/face-pixel coverage, paint
   brush/stroke/stroke-script coverage, renderer shader descriptor coverage,
   UI color parsing, and layout XML parse coverage.
 - `pano_cli inspect-image` reports PNG IHDR metadata as JSON and is covered by
@@ -103,8 +103,12 @@ Known local toolchain state:
   minimal PPI fixture.
 - `pp_assets_image_pixels_tests` decodes PNG payloads to RGBA8 and rejects
   corrupt image payloads.
-- `pano_cli load-project` creates a metadata-only `pp_document` projection with
-  per-layer frame counts and durations, and is covered by
+- `pp_document_ppi_import_tests` attaches decoded PPI dirty-face payloads to
+  `pp_document` layer/frame storage and rejects payloads outside document
+  layers.
+- `pano_cli load-project` creates a `pp_document` projection with per-layer
+  frame counts, durations, and decoded face-pixel payloads when present; the
+  metadata-only minimal fixture remains covered by
   `pano_cli_load_project_metadata_smoke`.
 - `pano_cli create-document` supports `--frames` and `--frame-duration-ms` and
   is covered by `pano_cli_create_animation_document_smoke`.