diff --git a/tests/test_propagate_valence.py b/tests/test_propagate_valence.py
new file mode 100644
index 000000000..5b8726250
--- /dev/null
+++ b/tests/test_propagate_valence.py
@@ -0,0 +1,81 @@
+"""Setr broker-confidence attenuation test.
+
+A->B->C chain. B is the broker. Low-confidence B should dampen propagation
+into C by ~9x compared to high-confidence B.
+"""
+from __future__ import annotations
+
+from aelfrice.models import (
+    BELIEF_FACTUAL,
+    EDGE_SUPPORTS,
+    LOCK_NONE,
+    Belief,
+    Edge,
+)
+from aelfrice.store import Store
+
+
+def _mk(bid: str, alpha: float, beta: float) -> Belief:
+    return Belief(
+        id=bid,
+        content=f"belief {bid}",
+        content_hash=f"h_{bid}",
+        alpha=alpha,
+        beta=beta,
+        type=BELIEF_FACTUAL,
+        lock_level=LOCK_NONE,
+        locked_at=None,
+        demotion_pressure=0,
+        created_at="2026-04-25T00:00:00Z",
+        last_retrieved_at=None,
+    )
+
+
+def _build_chain(broker_alpha: float, broker_beta: float) -> Store:
+    s = Store(":memory:")
+    # A is source; broker confidence at A doesn't matter for downstream
+    # because propagation multiplier uses dst broker, but give it neutral.
+    s.insert_belief(_mk("A", alpha=5.0, beta=5.0))
+    s.insert_belief(_mk("B", alpha=broker_alpha, beta=broker_beta))
+    # C is the terminus; give it neutral confidence so its broker factor
+    # is the same constant in both runs.
+    s.insert_belief(_mk("C", alpha=5.0, beta=5.0))
+    s.insert_edge(Edge(src="A", dst="B", type=EDGE_SUPPORTS, weight=1.0))
+    s.insert_edge(Edge(src="B", dst="C", type=EDGE_SUPPORTS, weight=1.0))
+    return s
+
+
+def test_low_broker_confidence_attenuates_vs_high() -> None:
+    # Low-confidence broker: alpha=1, beta=9 -> 0.1
+    low = _build_chain(broker_alpha=1.0, broker_beta=9.0)
+    # High-confidence broker: alpha=9, beta=1 -> 0.9
+    high = _build_chain(broker_alpha=9.0, broker_beta=1.0)
+
+    deltas_low = low.propagate_valence("A", valence=1.0, max_hops=3,
+                                       min_threshold=0.0001)
+    deltas_high = high.propagate_valence("A", valence=1.0, max_hops=3,
+                                         min_threshold=0.0001)
+
+    assert "C" in deltas_low, f"low: C missing, got {deltas_low}"
+    assert "C" in deltas_high, f"high: C missing, got {deltas_high}"
+
+    ratio = deltas_high["C"] / deltas_low["C"]
+    # Expected ratio: (0.9 / 0.1) = 9.0. Both runs share the same C broker
+    # factor, EDGE_VALENCE[SUPPORTS]=1.0, and source valence, so it cancels.
+    assert 8.5 < ratio < 9.5, f"expected ~9x ratio, got {ratio}"
+
+
+def test_propagate_returns_empty_for_isolated_source() -> None:
+    s = Store(":memory:")
+    s.insert_belief(_mk("solo", alpha=5.0, beta=5.0))
+    out = s.propagate_valence("solo", valence=1.0)
+    assert out == {}
+
+
+def test_propagate_respects_max_hops() -> None:
+    s = _build_chain(broker_alpha=5.0, broker_beta=5.0)
+    # max_hops=1 should reach B but not C.
+    out = s.propagate_valence("A", valence=1.0, max_hops=1,
+                              min_threshold=0.0001)
+    assert "B" in out
+    assert "C" not in out