diff --git a/trace_agg/tests/test_core.py b/trace_agg/tests/test_core.py
new file mode 100644
index 0000000..70884ac
--- /dev/null
+++ b/trace_agg/tests/test_core.py
@@ -0,0 +1,62 @@
+import pytest
+from trace_agg import core
+import math
+
+class TestCoreFunctions:
+    def test_compute_spearman_correlation_nominal(self):
+        retry_counts = [1, 2, 3, 4, 5]
+        window_durations = [2.0, 4.0, 6.0, 8.0, 10.0]
+        result = core.compute_spearman_correlation(retry_counts, window_durations)
+        assert isinstance(result, float)
+        # Spearman correlation for perfectly monotonic relationship should be 1.0
+        assert math.isclose(result, 1.0, rel_tol=1e-6)
+
+    def test_compute_spearman_correlation_inverse(self):
+        retry_counts = [1, 2, 3, 4, 5]
+        window_durations = [10.0, 8.0, 6.0, 4.0, 2.0]
+        result = core.compute_spearman_correlation(retry_counts, window_durations)
+        assert math.isclose(result, -1.0, rel_tol=1e-6)
+
+    def test_compute_spearman_correlation_invalid_inputs(self):
+        with pytest.raises(ValueError):
+            core.compute_spearman_correlation([1, 2], [3.0])
+
+    def test_calculate_edit_distance_identical_sequences(self):
+        seq = ["a", "b", "c"]
+        score = core.calculate_edit_distance(seq, seq)
+        assert score == 0.0
+
+    def test_calculate_edit_distance_completely_different(self):
+        seq1 = ["a", "b", "c"]
+        seq2 = ["x", "y", "z"]
+        score = core.calculate_edit_distance(seq1, seq2)
+        assert isinstance(score, float)
+        assert score == pytest.approx(3.0)
+
+    def test_calculate_edit_distance_partial_difference(self):
+        seq1 = ["a", "b", "c"]
+        seq2 = ["a", "x", "c"]
+        score = core.calculate_edit_distance(seq1, seq2)
+        assert score == pytest.approx(1.0)
+
+    def test_classify_runs_nominal(self):
+        run_data = [
+            {"run_id": "run1", "retry_free_reads": 10, "window_duration": 50.0},
+            {"run_id": "run2", "retry_free_reads": 20, "window_duration": 60.0}
+        ]
+        result = core.classify_runs(run_data)
+        assert isinstance(result, list)
+        assert len(result) == 2
+        for r in result:
+            assert set(r.keys()) == {"run_id", "step_stability_score", "correlation_coefficient"}
+            assert isinstance(r["run_id"], str)
+            assert isinstance(r["step_stability_score"], float)
+            assert isinstance(r["correlation_coefficient"], float)
+
+    def test_classify_runs_invalid_input_type(self):
+        with pytest.raises((AssertionError, TypeError, ValueError)):
+            core.classify_runs("not_a_list")
+
+    def test_classify_runs_edge_empty(self):
+        result = core.classify_runs([])
+        assert result == []
\ No newline at end of file