diff --git a/misc/codegen/generators/rusttestgen.py b/misc/codegen/generators/rusttestgen.py
index a44c5f2b241..b47d6b8725a 100644
--- a/misc/codegen/generators/rusttestgen.py
+++ b/misc/codegen/generators/rusttestgen.py
@@ -1,5 +1,7 @@
 import dataclasses
 import typing
+from collections.abc import Iterable
+
 import inflection
 
 from misc.codegen.loaders import schemaloader
@@ -27,6 +29,25 @@ class TestCode:
     function: Function | None = None
 
 
+def _get_code(doc: list[str]) -> list[str]:
+    adding_code = False
+    has_code = False
+    code = []
+    for line in doc:
+        match line, adding_code:
+            case ("```", _) | ("```rust", _):
+                adding_code = not adding_code
+                has_code = True
+            case _, False:
+                code.append(f"// {line}")
+            case _, True:
+                code.append(line)
+    assert not adding_code, "Unterminated code block in docstring:\n  " + "\n  ".join(doc)
+    if has_code:
+        return code
+    return []
+
+
 def generate(opts, renderer):
     assert opts.ql_test_output
     schema = schemaloader.load_file(opts.schema)
@@ -36,24 +57,18 @@ def generate(opts, renderer):
                          force=opts.force) as renderer:
         for cls in schema.classes.values():
             if (qlgen.should_skip_qltest(cls, schema.classes) or
-                    "rust_skip_test_from_doc" in cls.pragmas or
-                    not cls.doc):
+                    "rust_skip_doc_test" in cls.pragmas):
                 continue
-            code = []
-            adding_code = False
-            has_code = False
-            for line in cls.doc:
-                match line, adding_code:
-                    case ("```", _) | ("```rust", _):
-                        adding_code = not adding_code
-                        has_code = True
-                    case _, False:
-                        code.append(f"// {line}")
-                    case _, True:
-                        code.append(line)
-            if not has_code:
+            code = _get_code(cls.doc)
+            for p in schema.iter_properties(cls.name):
+                if "rust_skip_doc_test" in p.pragmas:
+                    continue
+                property_code = _get_code(p.description)
+                if property_code:
+                    code.append(f"// # {p.name}")
+                    code += property_code
+            if not code:
                 continue
-            assert not adding_code, "Unterminated code block in docstring: " + "\n".join(cls.doc)
             test_name = inflection.underscore(cls.name)
             signature = cls.pragmas.get("rust_doc_test_signature", "() -> ()")
             fn = signature and Function(f"test_{test_name}", signature)
diff --git a/misc/codegen/lib/schema.py b/misc/codegen/lib/schema.py
index bdaaba32c20..5ee769ca3f2 100644
--- a/misc/codegen/lib/schema.py
+++ b/misc/codegen/lib/schema.py
@@ -1,6 +1,7 @@
 """ schema format representation """
 import abc
 import typing
+from collections.abc import Iterable
 from dataclasses import dataclass, field
 from typing import List, Set, Union, Dict, Optional
 from enum import Enum, auto
@@ -143,6 +144,12 @@ class Schema:
     def null_class(self):
         return self.classes[self.null] if self.null else None
 
+    def iter_properties(self, cls: str) -> Iterable[Property]:
+        cls = self.classes[cls]
+        for b in cls.bases:
+            yield from self.iter_properties(b)
+        yield from cls.properties
+
 
 predicate_marker = object()