diff --git a/core/wren/src/wren/memory/schema_indexer.py b/core/wren/src/wren/memory/schema_indexer.py
index 7e28cd44fa..2e6ae62335 100644
--- a/core/wren/src/wren/memory/schema_indexer.py
+++ b/core/wren/src/wren/memory/schema_indexer.py
@@ -51,13 +51,16 @@ def describe_schema(manifest: dict) -> str:
         lines.append("")
 
     for model in manifest.get("models", []):
-        _describe_model(model, lines)
+        if isinstance(model, dict) and model.get("name"):
+            _describe_model(model, lines)
 
     for rel in manifest.get("relationships", []):
-        _describe_relationship(rel, lines)
+        if isinstance(rel, dict):
+            _describe_relationship(rel, lines)
 
     for view in manifest.get("views", []):
-        _describe_view(view, lines)
+        if isinstance(view, dict) and view.get("name"):
+            _describe_view(view, lines)
 
     cubes = manifest.get("cubes", []) or []
     if isinstance(cubes, list):
@@ -87,7 +90,9 @@ def _describe_model(model: dict, lines: list[str]) -> None:
     if data_scope:
         lines.append(f"  Data scope: {data_scope}")
 
-    cols = model.get("columns", [])
+    cols = [
+        c for c in (model.get("columns") or []) if isinstance(c, dict) and c.get("name")
+    ]
     if cols:
         lines.append("  Columns:")
         for col in cols:
@@ -96,7 +101,9 @@ def _describe_model(model: dict, lines: list[str]) -> None:
 
 
 def _describe_column(col: dict, lines: list[str]) -> None:
-    name = col["name"]
+    name = col.get("name")
+    if not name:
+        return
     dtype = col.get("type", "?")
     parts = [f"    - {name} ({dtype})"]
 
@@ -229,14 +236,22 @@ def extract_schema_items(manifest: dict) -> list[dict]:
     items: list[dict] = []
 
     for model in manifest.get("models", []):
+        if not isinstance(model, dict) or not model.get("name"):
+            continue
         items.append(_model_record(model, mdl_h, now))
-        for col in model.get("columns", []):
+        for col in model.get("columns", []) or []:
+            if not isinstance(col, dict) or not col.get("name"):
+                continue
             items.append(_column_record(col, model["name"], mdl_h, now))
 
     for rel in manifest.get("relationships", []):
+        if not isinstance(rel, dict):
+            continue
         items.append(_relationship_record(rel, mdl_h, now))
 
     for view in manifest.get("views", []):
+        if not isinstance(view, dict):
+            continue
         items.append(_view_record(view, mdl_h, now))
 
     cubes = manifest.get("cubes", []) or []
@@ -264,7 +279,9 @@ def extract_schema_items(manifest: dict) -> list[dict]:
 
 def _model_record(model: dict, mdl_h: str, now: datetime) -> dict:
     name = model["name"]
-    cols = model.get("columns", [])
+    cols = [
+        c for c in (model.get("columns") or []) if isinstance(c, dict) and c.get("name")
+    ]
     col_summaries = ", ".join(f"{c['name']} ({c.get('type', '?')})" for c in cols[:20])
     pk = model.get("primaryKey") or ""
 
@@ -297,6 +314,7 @@ def _model_record(model: dict, mdl_h: str, now: datetime) -> dict:
 
 
 def _column_record(col: dict, model_name: str, mdl_h: str, now: datetime) -> dict:
+    # Caller (extract_schema_items) already guarantees a truthy name.
     name = col["name"]
     dtype = col.get("type", "")
     expr = col.get("expression") or None
diff --git a/core/wren/tests/unit/test_schema_indexer_malformed_columns.py b/core/wren/tests/unit/test_schema_indexer_malformed_columns.py
new file mode 100644
index 0000000000..0209af289f
--- /dev/null
+++ b/core/wren/tests/unit/test_schema_indexer_malformed_columns.py
@@ -0,0 +1,88 @@
+"""schema_indexer must skip non-dict / nameless columns."""
+
+from __future__ import annotations
+
+from wren.memory.schema_indexer import describe_schema, extract_schema_items
+
+
+def test_describe_schema_skips_non_dict_columns():
+    text = describe_schema(
+        {
+            "models": [
+                {
+                    "name": "orders",
+                    "columns": [None, {"name": "amount", "type": "int"}, {"type": "x"}],
+                }
+            ]
+        }
+    )
+    assert "amount" in text
+    # Exactly one column entry rendered (the malformed ones are dropped).
+    assert text.count("    - ") == 1
+
+
+def test_describe_schema_omits_dangling_columns_header():
+    # When every column is malformed, no dangling "  Columns:" header (589e0bb).
+    text = describe_schema(
+        {"models": [{"name": "orders", "columns": [None, {"type": "x"}]}]}
+    )
+    assert "Columns:" not in text
+    assert "### Model: orders" in text
+
+
+def test_describe_schema_skips_empty_string_name_and_null_columns():
+    text = describe_schema(
+        {
+            "models": [
+                {"name": "a", "columns": [{"name": "", "type": "int"}]},
+                {"name": "b", "columns": None},
+            ]
+        }
+    )
+    assert "### Model: a" in text
+    assert "### Model: b" in text
+    assert "    - " not in text
+
+
+def test_describe_schema_skips_null_model_relationship_view_slots():
+    # Null slots in hand-edited manifests must not crash.
+    text = describe_schema({"models": [None], "relationships": [None], "views": [None]})
+    assert isinstance(text, str)
+
+
+def test_extract_schema_items_skips_null_model_slots():
+    items = extract_schema_items({"models": [None, {"columns": []}]})
+    assert items == []
+
+
+def test_extract_schema_items_skips_scalar_and_empty_name_columns():
+    items = extract_schema_items(
+        {
+            "models": [
+                {
+                    "name": "orders",
+                    "columns": ["amount", {"name": "", "type": "int"}],
+                }
+            ]
+        }
+    )
+    col_items = [i for i in items if i["item_type"] == "column"]
+    assert col_items == []
+
+
+def test_extract_schema_items_skips_non_dict_columns():
+    items = extract_schema_items(
+        {
+            "models": [
+                {
+                    "name": "orders",
+                    "columns": [None, {"name": "amount", "type": "int"}, {"type": "x"}],
+                }
+            ]
+        }
+    )
+    col_items = [i for i in items if i["item_type"] == "column"]
+    assert len(col_items) == 1
+    assert col_items[0]["item_name"] == "amount"
+    model_items = [i for i in items if i["item_type"] == "model"]
+    assert "amount" in model_items[0]["text"]