Add passive tabletop lab protocol

core/cli.py

@@ -133,6 +133,7 @@ _TEST_SUITES: dict[str, tuple[str, ...]] = {
         "tests/test_environment_falsification.py",
         "tests/test_environment_falsification_eval_cli.py",
         "tests/test_witness_log_importer.py",
+        "tests/test_tabletop_lab_protocol.py",
         "tests/test_sensorium_eval_cli.py",
         "tests/test_efferent_gate.py",
     ),

docs/lab/tabletop_falsification_lab.md

@@ -0,0 +1,43 @@
+# Passive Tabletop Falsification Lab Protocol
+
+## Purpose
+
+The tabletop lab is an offline evidence source for ADR-0211 falsification. It is
+not a live robotics stack and it does not authorize CORE-controlled actuation.
+
+## Hardware Target
+
+- Event camera for sparse brightness-change deltas.
+- RGB/depth camera for reference visual frames.
+- AprilTag calibration board and object tags for pose witness evidence.
+- IMU/contact/proprioception logger for afferent sensorimotor evidence.
+
+## First Lab Mode
+
+Only passive, human-moved objects are allowed. The lab records logs, imports
+them offline, compiles afferent units, builds `ObservationFrame`s, and runs
+falsification scenarios.
+
+```text
+passive environment movement
+  -> offline witness logs
+  -> deterministic import
+  -> afferent compilers
+  -> ObservationFrame sequence
+  -> FalsificationScenario report
+```
+
+## Acceptance Gates
+
+- The same captured log imports to the same frame-sequence hash.
+- AprilTag/GTSAM-style pose outputs remain witness evidence only.
+- Event-vision and sensorimotor deltas remain compiler-native.
+- No raw camera frame, event stream, PCM, trajectory, or actuator trace enters
+  `ObservationFrame` or falsification traces.
+- No CORE motor command path is mounted.
+
+## Deferred
+
+Repeated capture statistics, hardware-noise envelopes, live device readers,
+robot actuation, and motor decoder work require later ADRs. Exact replay remains
+the first gate.

tests/test_tabletop_lab_protocol.py

@@ -0,0 +1,14 @@
+from __future__ import annotations
+
+from pathlib import Path
+
+
+def test_tabletop_lab_protocol_is_passive_and_trace_safe():
+    root = Path(__file__).resolve().parents[1]
+    doc = root / "docs" / "lab" / "tabletop_falsification_lab.md"
+    assert doc.exists()
+    text = doc.read_text(encoding="utf-8")
+    assert "Only passive, human-moved objects are allowed" in text
+    assert "No CORE motor command path is mounted" in text
+    assert "No raw camera frame, event stream, PCM, trajectory, or actuator trace enters" in text
+    assert "Exact replay remains" in text