fix: respect zero policy sample thresholds

src/flightdeck/ledger.py

@@ -183,9 +183,11 @@ def diff_releases(
     candidate_rollup = compute_rollup(candidate_events, candidate_pricing_table)
 
     # Confidence (policy can override thresholds; otherwise take config defaults)
-    min_candidate_runs = policy.min_candidate_runs or cfg.diff.min_candidate_runs
-    min_baseline_runs = policy.min_baseline_runs or cfg.diff.min_baseline_runs
-    min_low_runs = policy.min_low_runs or cfg.diff.min_low_runs
+    min_candidate_runs = (
+        policy.min_candidate_runs if policy.min_candidate_runs is not None else cfg.diff.min_candidate_runs
+    )
+    min_baseline_runs = policy.min_baseline_runs if policy.min_baseline_runs is not None else cfg.diff.min_baseline_runs
+    min_low_runs = policy.min_low_runs if policy.min_low_runs is not None else cfg.diff.min_low_runs
 
     label = confidence_label(
         baseline_rollup.runs,

tests/test_ledger.py

@@ -89,3 +89,27 @@ def test_diff_releases_rejects_mixed_agents_within_side() -> None:
             candidate_pricing_table=table,
             window="7d",
         )
+
+
+def test_diff_releases_respects_zero_policy_sample_thresholds() -> None:
+    cfg = WorkspaceConfig()
+    policy = Policy(
+        min_baseline_runs=0,
+        min_candidate_runs=0,
+        min_low_runs=0,
+        require_high_diff_confidence=True,
+    )
+    table = _pricing_table()
+
+    result = diff_releases(
+        cfg=cfg,
+        policy=policy,
+        baseline_events=[],
+        candidate_events=[],
+        baseline_pricing_table=table,
+        candidate_pricing_table=table,
+        window="7d",
+    )
+
+    assert result.confidence == "HIGH"
+    assert result.policy.passed