@@ -14,38 +14,38 @@ means a green pipeline.
 | `python3 tests/test_density.py` | nothing | backend gap-fill and edge-extension rules |
 | `node tests/test_density_rail_parity.mjs` | Node | live-rail edge-gap parity with the backend, including gaps larger than 25% of HS |
 | `python3 tests/test_instrument_state.py` | nothing | Y / N / unanswered persistence, CSV export, re-archive and contradiction validation |
-| `python3 tests/test_stage6_integration.py` | nothing | cumulative site identity, legacy migration, staged first archive, recovery, re-archive renaming, attachment preservation, duplicate-ID and instrument-state behavior |
-| `python3 tests/test_stage8_photo_manifest.py` | nothing | expected-photo schema/migration, archive registration, recovery, explicit cancellation, limits, queue-ID identity and idempotent server upload |
-| `python3 tests/test_stage9_attachment_consistency.py` | nothing | staged publication, crash recovery, DB-failure compensation, missing/orphan reconciliation, stale-temp cleanup, safe deletion, ownership, and Stage 8-to-9 migration |
-| `python3 tests/test_stage10_saved_pits.py` | nothing | owner-scoped Saved Pits search, campaign/date filters, sort order, pagination, recovery separation, status counts, API validation, and indexes |
-| `python3 tests/test_stage11_workspace.py` | nothing | owner-scoped workspace summary, recovery/photo counts, API identity boundary, and assembled initial view |
-| `python3 tests/test_stage12_security.py` | nothing | fail-closed identity parsing, owner-bound CSRF tokens, configuration requirements, and rate limiting |
-| `python3 tests/test_stage12_ops.py` | nothing | consistent backup, manifest verification, traversal/tamper rejection, restore, and rollback copies |
-| `python3 tests/test_stage12_storage_lifecycle.py` | nothing | shared archive/attachment lock, stale-folder rejection, degraded-lock warnings, and rename durability |
-| `python3 tests/test_stage13_ui.py` | nothing | offline font independence, workspace hierarchy, section subtitles, responsive/contrast/print modes, and explicit attachment-state styling |
-| `python3 tests/test_stage14_transfer.py` | nothing | installation/revision identities, one-way bundles, dry-run, owner mapping, idempotency, fast-forward updates, multiple field databases, attachment transfer, recovery resume, tamper rejection, conflicts, and audit rows |
-| `python3 tests/test_resource_stage1_downloads.py` | nothing | disk-backed download ZIP contents, normal/interrupted cleanup, and startup stale-file recovery |
-| `python3 tests/test_resource_stage2_uploads.py` | nothing | bounded upload-stream staging, incremental digest/header checks, rejection cleanup, and startup stale-file recovery |
-| `python3 tests/test_resource_stage3_heic.py` | Pillow | HEIC conversion semaphore, failure-safe permit release, disk-backed JPEG staging, crash cleanup, and conversion-before-storage-lock ordering |
-| `python3 tests/test_resource_stage4_profiles.py` | matplotlib | profile render semaphore, failure-safe permit release, single-build PNG+PDF equivalence, and 300-DPI guardrail |
-| `python3 tests/test_resource_stage5_threads.py` | nothing | eight-thread shared-server default, four-thread override, Waitress thread wiring, and startup concurrency diagnostics |
-| `python3 tests/test_resource_stage6_resilience.py` | nothing | abrupt-process-kill recovery for download/upload scratch files and process-local HEIC/profile permits |
-| `python3 tests/test_resource_stage6_live.py` | flask, waitress, matplotlib | boots the real eight-thread Waitress server, verifies startup scratch cleanup, runs two profile requests concurrently, and checks health traffic remains responsive |
-| `python3 tests/test_resource_stage7_sizing.py` | nothing | Stage 7 sizing-policy PASS/PROVISIONAL/FAIL boundaries for peak RSS, host headroom, HEIC mode, soak stability, swap, latency, and cleanup |
+| `python3 tests/test_archive_integration.py` | nothing | cumulative site identity, legacy migration, staged first archive, recovery, re-archive renaming, attachment preservation, duplicate-ID and instrument-state behavior |
+| `python3 tests/test_photo_manifest.py` | nothing | expected-photo schema/migration, archive registration, recovery, explicit cancellation, limits, queue-ID identity and idempotent server upload |
+| `python3 tests/test_attachment_consistency.py` | nothing | staged publication, crash recovery, DB-failure compensation, missing/orphan reconciliation, stale-temp cleanup, safe deletion, ownership, and legacy attachment-schema migration |
+| `python3 tests/test_saved_pits.py` | nothing | owner-scoped Saved Pits search, campaign/date filters, sort order, pagination, recovery separation, status counts, API validation, and indexes |
+| `python3 tests/test_workspace.py` | nothing | owner-scoped workspace summary, recovery/photo counts, API identity boundary, and assembled initial view |
+| `python3 tests/test_security_core.py` | nothing | fail-closed identity parsing, owner-bound CSRF tokens, configuration requirements, and rate limiting |
+| `python3 tests/test_backup_restore.py` | nothing | consistent backup, manifest verification, traversal/tamper rejection, restore, and rollback copies |
+| `python3 tests/test_storage_lifecycle.py` | nothing | shared archive/attachment lock, stale-folder rejection, degraded-lock warnings, and rename durability |
+| `python3 tests/test_interface_contract.py` | nothing | offline font independence, workspace hierarchy, section subtitles, responsive/contrast/print modes, and explicit attachment-state styling |
+| `python3 tests/test_field_transfer.py` | nothing | installation/revision identities, one-way bundles, dry-run, owner mapping, idempotency, fast-forward updates, multiple field databases, attachment transfer, recovery resume, tamper rejection, conflicts, and audit rows |
+| `python3 tests/test_resource_downloads.py` | nothing | disk-backed download ZIP contents, normal/interrupted cleanup, and startup stale-file recovery |
+| `python3 tests/test_resource_uploads.py` | nothing | bounded upload-stream staging, incremental digest/header checks, rejection cleanup, and startup stale-file recovery |
+| `python3 tests/test_resource_heic.py` | Pillow | HEIC conversion semaphore, failure-safe permit release, disk-backed JPEG staging, crash cleanup, and conversion-before-storage-lock ordering |
+| `python3 tests/test_resource_profiles.py` | matplotlib | profile render semaphore, failure-safe permit release, single-build PNG+PDF equivalence, and 300-DPI guardrail |
+| `python3 tests/test_resource_threads.py` | nothing | eight-thread shared-server default, four-thread override, Waitress thread wiring, and startup concurrency diagnostics |
+| `python3 tests/test_resource_resilience.py` | nothing | abrupt-process-kill recovery for download/upload scratch files and process-local HEIC/profile permits |
+| `python3 tests/test_resource_live_server.py` | flask, waitress, matplotlib | boots the real eight-thread Waitress server, verifies startup scratch cleanup, runs two profile requests concurrently, and checks health traffic remains responsive |
+| `python3 tests/test_resource_sizing.py` | nothing | server-sizing PASS/PROVISIONAL/FAIL boundaries for peak RSS, host headroom, HEIC mode, soak stability, swap, latency, and cleanup |
 | `python3 tests/test_weather_multiselect.py` | nothing | canonical weather arrays, legacy scalar compatibility, validation, normalized reporting columns, and CSV export |
-| `node tests/test_stage13_ui.mjs` | nothing | automatic accessible field names and non-invasive populated-field presentation state |
+| `node tests/test_interface_ui.mjs` | nothing | automatic accessible field names and non-invasive populated-field presentation state |
 | `node tests/test_table_enter_navigation.mjs` | nothing | Enter moves vertically through existing Temperature, Density, LWC, and Stratigraphy rows without adding rows or submitting |
 | `node tests/test_clearable_radios.mjs` | nothing | optional native-radio clear action, accessible naming, idempotent setup, and no-scroll focus return |
 | `node tests/test_weather_multiselect.mjs` | nothing | native checkbox coexistence, precipitation None exclusivity, array collection, and legacy/new restore behavior |
-| `python3 tests/test_stage12_flask_security.py` | flask, matplotlib | real route-level CSRF, headers, health/readiness, size limits, generic errors, and Alice/Bob isolation |
-| `node tests/test_stage12_csrf_ui.mjs` | nothing | browser API helper sends CSRF only on unsafe requests |
+| `python3 tests/test_flask_security.py` | flask, matplotlib | real route-level CSRF, headers, health/readiness, size limits, generic errors, and Alice/Bob isolation |
+| `node tests/test_csrf_ui.mjs` | nothing | browser API helper sends CSRF only on unsafe requests |
 | `node tests/test_instrument_ui.mjs` | nothing | lightweight real-function checks for collect() and Y/N/unanswered button state |
 | `node tests/test_record_workflow_ui.mjs` | nothing | lightweight execution of edit mode, pending recovery state, Start New Pit, and site_id-based loading |
 | `node tests/test_attachment_outbox.mjs` | nothing | durable queue persistence, restore, draft-to-site binding, quota failure, duplicate selection, and explicit discard |
 | `node tests/test_attachment_flush.mjs` | nothing | failed upload retention, confirmed upload cleanup, duplicate confirmation, and network retry state |
 | `node tests/test_attachment_manifest_ui.mjs` | nothing | archive manifest handoff, cancellation safety, server-only expectations, and stored-attachment deletion controls |
-| `node tests/test_stage10_saved_pits_ui.mjs` | nothing | finder query construction, status cards, campaign facets, recovery separation, and load-more accumulation |
-| `node tests/test_stage11_workspace_ui.mjs` | nothing | workspace navigation, current-record awareness, recent/recovery actions, finder-state preservation, and all-context photo-queue summary |
+| `node tests/test_saved_pits_ui.mjs` | nothing | finder query construction, status cards, campaign facets, recovery separation, and load-more accumulation |
+| `node tests/test_workspace_ui.mjs` | nothing | workspace navigation, current-record awareness, recent/recovery actions, finder-state preservation, and all-context photo-queue summary |
 | `python3 tests/test_plot.py` | matplotlib | profile figure geometry |
 | `python3 tests/test_smoke.py` | flask, matplotlib | end-to-end: archive -> load -> download, the seven CSVs + PNG, attachment limits, the pit-sheet PDF/image rule |
 | `node tests/test_coords.mjs` | nothing | UTM/WGS84 converter vs PROJ ground truth |
@@ -57,17 +57,17 @@ runtimes — it is not a pure-JS suite.
 ## Server resource qualification
 
 The normal suite keeps tests bounded for CI. Before sizing or changing a shared
-server, run the Stage 6 resource harness on that host after installing the locked
+server, run the server-resource harness on that host after installing the locked
 Python dependencies:
 
 ```bash
-python3 tests/benchmark_resource_stage6.py --qualification --output stage6.json
+python3 tests/benchmark_server_resources.py --qualification --output server-resources.json
 ```
 
 For a multi-hour stability check, add a wall-clock soak, for example:
 
 ```bash
-python3 tests/benchmark_resource_stage6.py --qualification --soak-minutes 180 --output stage6-soak.json
+python3 tests/benchmark_server_resources.py --qualification --soak-minutes 180 --output server-resource-soak.json
 ```
 
 The harness reports peak/current RSS, host total RAM, minimum `MemAvailable`,
@@ -77,10 +77,10 @@ codec or the clearly labelled decoded-image proxy. On Linux it uses `/proc` if
 `psutil` is not installed; on other platforms install `psutil` only for
 qualification. `psutil` is intentionally not a CryoPit runtime dependency.
 
-Evaluate a completed target-host soak against the Stage 7 3.5 GiB policy with:
+Evaluate a completed target-host soak against the 3.5 GiB server-sizing policy with:
 
 ```bash
-python3 tests/evaluate_resource_stage7.py stage6-soak.json --ram-gib 3.5
+python3 tests/evaluate_server_sizing.py server-resource-soak.json --ram-gib 3.5
 ```
 
 The evaluator returns `PASS`, `PROVISIONAL`, or `FAIL`. It deliberately refuses
@@ -102,7 +102,7 @@ checks that it opened.
 
 CryoPit supports Python 3.11 and newer. CI installs `requirements.lock` across the supported Python matrix and also runs the full dependency-backed suite in a Python 3.11 Conda environment. A separate floating-dependency job installs `requirements.txt` to detect future compatibility drift.
 
-The scientific, cumulative integration, operations, security-core, and lightweight JavaScript suites do not require Flask or jsdom. The end-to-end and route-isolation suites require the Python dependencies in `requirements.lock`; the DOM suite requires `npm ci`. A syntax-only pass is not reported as a runtime pass when those dependencies are unavailable. A Stage 14 release candidate is not production-ready until the full dependency-backed suite passes in the target build environment. The Stage 13 interface also receives a real-browser layout/runtime check at representative desktop, tablet, and mobile widths; screenshot review supplements but does not replace the DOM assertions.
+The scientific, cumulative integration, operations, security-core, and lightweight JavaScript suites do not require Flask or jsdom. The end-to-end and route-isolation suites require the Python dependencies in `requirements.lock`; the DOM suite requires `npm ci`. A syntax-only pass is not reported as a runtime pass when those dependencies are unavailable. A release candidate is not production-ready until the full dependency-backed suite passes in the target build environment. The interface also receives a real-browser layout/runtime check at representative desktop, tablet, and mobile widths; screenshot review supplements but does not replace the DOM assertions.
 
 ### DOM placeholder guard
 
 
@@ -1,13 +1,13 @@
-"""Reproducible CryoPit Stage 6 server-resource qualification harness.
+"""Reproducible CryoPit server-resource qualification harness.
 
 This is intentionally not part of the normal CI suite: qualification loads are
 large and machine-specific. Run it on the deployment host after installing
-``requirements.lock``. It exercises the exact Stage 1/2/3/4 resource paths in
+``requirements.lock``. It exercises the production download/upload/HEIC/profile resource paths in
 one process and emits JSON suitable for attaching to an infrastructure ticket.
 
 Examples:
-  python tests/benchmark_resource_stage6.py --quick
-  python tests/benchmark_resource_stage6.py --qualification --output stage6.json
+  python tests/benchmark_server_resources.py --quick
+  python tests/benchmark_server_resources.py --qualification --output server-resources.json
 """
 from __future__ import annotations
 
@@ -41,7 +41,7 @@ def __init__(self):
             self._proc = None
         else:
             raise SystemExit(
-                "Stage 6 benchmark needs psutil on this platform; install it only "
+                "Server-resource benchmark needs psutil on this platform; install it only "
                 "for qualification (it is not a CryoPit runtime dependency)."
             )
 
@@ -192,7 +192,7 @@ def main():
         image = Image.new("RGB", (width, height), (135, 145, 155))
         image.save(source, format="JPEG", quality=95)
         image.close()
-        # Let the Stage 3 file-to-file helper exercise the same Pillow decode /
+        # Let the file-to-file helper exercise the same Pillow decode /
         # encode path without claiming this is HEIC codec performance.
         import types
         sys.modules["pillow_heif"] = types.SimpleNamespace(register_heif_opener=lambda: None)
 
@@ -1,11 +1,11 @@
-"""Evaluate a Stage 6 qualification report against CryoPit's Stage 7 RAM policy.
+"""Evaluate a server-resource qualification report against CryoPit's RAM policy.
 
 This does not benchmark the machine. It evaluates JSON produced by
-``tests/benchmark_resource_stage6.py`` and reports whether a planned memory
+``tests/benchmark_server_resources.py`` and reports whether a planned memory
 allocation has been demonstrated on a representative host.
 
 Example:
-    python tests/evaluate_resource_stage7.py stage6-soak.json --ram-gib 3.5
+    python tests/evaluate_server_sizing.py server-resource-soak.json --ram-gib 3.5
 """
 from __future__ import annotations
 
@@ -53,7 +53,7 @@ def evaluate(report: dict, ram_gib: float) -> dict:
         failures.append("peak RSS is missing")
     elif peak > peak_limit:
         failures.append(
-            f"CryoPit peak RSS {peak:.0f} MiB exceeds the Stage 7 planning ceiling "
+            f"CryoPit peak RSS {peak:.0f} MiB exceeds the server planning ceiling "
             f"of {peak_limit:.0f} MiB ({APP_PEAK_FRACTION:.0%} of target RAM)"
         )
     else:
@@ -68,7 +68,7 @@ def evaluate(report: dict, ram_gib: float) -> dict:
     elif system_total is not None and system_total <= ram_mib * (1.0 + TARGET_HOST_TOLERANCE):
         if min_available < available_floor:
             failures.append(
-                f"minimum MemAvailable {min_available:.0f} MiB is below the Stage 7 "
+                f"minimum MemAvailable {min_available:.0f} MiB is below the server headroom "
                 f"floor of {available_floor:.0f} MiB"
             )
         else:
@@ -157,7 +157,7 @@ def evaluate(report: dict, ram_gib: float) -> dict:
 
 def main() -> int:
     parser = argparse.ArgumentParser()
-    parser.add_argument("report", help="Stage 6 qualification JSON")
+    parser.add_argument("report", help="server-resource qualification JSON")
     parser.add_argument("--ram-gib", type=float, default=3.5,
                         help="planned server RAM in GiB (default: 3.5)")
     parser.add_argument("--output", help="optional path for evaluator JSON")
 
@@ -1,4 +1,4 @@
-"""Cumulative Stage 1-5 integration tests without Flask.
+"""Cumulative archive integration tests without Flask.
 
 Covers immutable site identity, legacy migration, recoverable first archive,
 recoverable re-archive folder renaming, in-place attachment preservation, edit
@@ -353,5 +353,5 @@ def test_archive_retry_returns_pending_conflict_instead_of_raising():
     for test in tests:
         test()
         print("PASS", test.__name__)
-    print(f"{len(tests)} Stage 6 integration tests passed")
+    print(f"{len(tests)} archive integration tests passed")
 
@@ -1,4 +1,4 @@
-"""Stage 9 attachment filesystem/SQLite compensation and recovery tests."""
+"""Attachment filesystem/SQLite compensation and recovery tests."""
 from __future__ import annotations
 
 import hashlib
@@ -140,9 +140,9 @@ def prepare_staged(site_id, item, data, filename="reserved.jpg"):
     return folder, staged, rel
 
 
-def test_stage8_database_migrates_before_new_index_is_created():
+def test_legacy_database_migrates_before_new_index_is_created():
     # CREATE TABLE IF NOT EXISTS cannot add columns to an existing table, so
-    # indexes involving Stage 9 columns must be created by _migrate afterward.
+    # indexes involving later attachment columns must be created by _migrate afterward.
     db_path = Path(os.environ["CRYOPIT_DB_PATH"])
     exports = Path(os.environ["CRYOPIT_EXPORT_DIR"])
     if db_path.exists(): db_path.unlink()
@@ -162,7 +162,7 @@ def test_stage8_database_migrates_before_new_index_is_created():
         ):
             legacy_sql = legacy_sql.replace(line, "")
         # Removing the final attachment column leaves the preceding timestamp
-        # with a comma; restore the Stage 8 closing form.
+        # with a comma; restore the legacy closing form.
         legacy_sql = legacy_sql.replace(
             "    uploaded_at TEXT DEFAULT (datetime('now')),\n);",
             "    uploaded_at TEXT DEFAULT (datetime('now'))\n);")
@@ -365,4 +365,4 @@ def test_delete_enforces_owner():
     tests = sorted((n, f) for n, f in globals().items() if n.startswith("test_") and callable(f))
     for name, fn in tests:
         fn(); print("PASS", name)
-    print(f"{len(tests)} Stage 9 attachment-consistency tests passed")
+    print(f"{len(tests)} attachment-consistency tests passed")