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_foxglove_bridge.py

@@ -67,3 +67,20 @@ def test_metrics_distinguish_preview_drops_and_pipeline_latency() -> None:
     assert snapshot["last_point_count"] == 42
     assert snapshot["mqtt_to_publish_ms"] == 12.346
     assert snapshot["decode_publish_ms"] == 1.234
+
+
+def test_metrics_report_live_ai_latency_rate_staleness_and_drops() -> None:
+    metrics = BridgeMetrics()
+    metrics.perception_dropped()
+    metrics.published_perception(
+        captured_at_epoch_ns=1_000_000_000,
+        published_at_epoch_ns=1_125_000_000,
+        published_monotonic_ns=2_000_000_000,
+    )
+
+    snapshot = metrics.snapshot()
+
+    assert snapshot["perception_frames"] == 1
+    assert snapshot["perception_dropped"] == 1
+    assert snapshot["perception_end_to_end_ms"] == 125.0
+    assert snapshot["perception_end_to_end_p95_ms"] == 125.0