diff --git a/stock/app/holdings_intel.py b/stock/app/holdings_intel.py
index f760026..49a3b0c 100644
--- a/stock/app/holdings_intel.py
+++ b/stock/app/holdings_intel.py
@@ -152,7 +152,7 @@ def market_events(ticker: str, ticker_prices: "pd.DataFrame",
         last_vol = vol.iloc[-1]
         if mu > 0 and (
             (sd and (last_vol - mu) / sd >= p["vol_z"])
-            or (not sd and last_vol >= mu * p["vol_z"])  # sd=0 fallback: plain ratio
+            or (not sd and last_vol >= mu * p["vol_z"])  # sd=0 (평탄 기준선): vol_z를 Z-score가 아닌 단순 배수로 사용
         ):
             z_txt = f"{(last_vol - mu) / sd:.1f}" if sd else f"ratio={last_vol / mu:.1f}x"
             events.append({
diff --git a/stock/app/test_holdings_intel.py b/stock/app/test_holdings_intel.py
index 455be8e..30dcee7 100644
--- a/stock/app/test_holdings_intel.py
+++ b/stock/app/test_holdings_intel.py
@@ -226,3 +226,50 @@ def test_portfolio_health():
     assert h["positions"] == 2
     assert 0 <= h["max_weight"] <= 1.0
     assert "total_eval" in h and "total_pnl" in h and "cash_ratio" in h
+
+
+def test_market_events_volume_surge_zscore_path():
+    # 변동 있는 기준선 → Z-score 경로(sd>0) 검증 (sd=0 fallback 아님)
+    import random as _r
+    _r.seed(1)
+    base_vols = [1000 + _r.randint(-50, 50) for _ in range(30)]
+    closes = [1000] * 30 + [1010]
+    vols = base_vols + [max(base_vols) * 10]  # 마지막날 큰 급증
+    df = _ticker_prices(closes, vols)
+    evts = hi.market_events("005930", df, None, DEFAULT_EVENT)
+    assert any(e["type"] == "volume_surge" for e in evts)
+
+
+def test_market_events_foreign_selling():
+    closes = [1000] * 5
+    df = _ticker_prices(closes)
+    import datetime as _dt
+    base = _dt.date(2025, 1, 1)
+    flow = pd.DataFrame({
+        "ticker": ["005930"] * 5,
+        "date": [(base + _dt.timedelta(days=i)).isoformat() for i in range(5)],
+        "foreign_net": [100, 50, -10, -20, -30],   # 최근 3일 연속 순매도
+        "institution_net": [0] * 5,
+    })
+    evts = hi.market_events("005930", df, flow, DEFAULT_EVENT)
+    assert any(e["type"] == "foreign_selling" for e in evts)
+
+
+def test_news_issues_severity_high_boundary(monkeypatch):
+    monkeypatch.setattr(hi, "_news_sentiment_map", lambda date: {
+        "005930": {"score_raw": -0.6, "news_count": 5}})   # 정확히 high 경계
+    issues = hi.news_issues(["005930"], date="2026-05-29", use_llm=False)
+    assert issues["005930"][0]["severity"] == "high"
+
+
+def test_portfolio_health_empty_and_zero():
+    # 빈 포트 → 0/빈값, 크래시 없음
+    h0 = hi.portfolio_health([], total_cash=0)
+    assert h0["positions"] == 0
+    assert h0["max_weight"] == 0.0
+    assert h0["total_pnl_rate"] == 0.0
+    assert h0["cash_ratio"] == 0.0
+    # total_buy=0 (avg_price 0) → div-by-zero 없이 0.0
+    h1 = hi.portfolio_health([{"ticker": "X", "quantity": 1, "avg_price": 0,
+                               "current_price": 0, "is_krx": True}], total_cash=0)
+    assert h1["total_pnl_rate"] == 0.0