feat(perception): integrate calibrated operator pipeline

Add calibrated K1 projection, recorded and near-live perception qualification, unified Rerun operator layers, bounded replay admission, audited viewer controls, worker experiments, and lab evidence.

tests/test_xgrids_ap_activation.py

@@ -0,0 +1,35 @@
+from k1link.device_plugins.xgrids_k1.ble.ap_activation import (
+    COMMAND_OFFSET,
+    ENABLE_AP_COMMAND,
+    FRAME_LENGTH,
+    build_ap_activation_frame,
+    is_ap_ready_status,
+)
+from k1link.device_plugins.xgrids_k1.ble.wifi_provisioning import WifiStatus
+
+
+def test_build_ap_activation_frame_matches_reviewed_lixelgo_layout() -> None:
+    frame = build_ap_activation_frame()
+
+    assert len(frame) == FRAME_LENGTH == 100
+    assert frame[COMMAND_OFFSET] == ENABLE_AP_COMMAND == 1
+    assert frame[:COMMAND_OFFSET] == bytes(COMMAND_OFFSET)
+
+
+def _status(*, reserved: int) -> WifiStatus:
+    return {
+        "value_length": 52,
+        "mode": "WIFI_AP",
+        "ipv4": "192.168.56.1",
+        "status_code": 1,
+        "reserved": reserved,
+        "trailer_hex": "",
+    }
+
+
+def test_ap_control_mode_without_ready_flag_is_not_ready() -> None:
+    assert not is_ap_ready_status(_status(reserved=0))
+
+
+def test_ap_ready_requires_the_reviewed_byte_51_flag() -> None:
+    assert is_ap_ready_status(_status(reserved=1))