fix(http-client-python): prevent external entity parsing

.chronus/changes/main-2025-8-17-12-23-40.md

@@ -0,0 +1,7 @@
+---
+changeKind: fix
+packages:
+  - "@typespec/http-client-python"
+---
+
+Explicitly use element tree parser with resolve_entities set to False

packages/http-client-python/generator/pygen/codegen/templates/model_base.py.jinja2

@@ -17,6 +17,7 @@ import enum
 import email.utils
 from datetime import datetime, date, time, timedelta, timezone
 from json import JSONEncoder
+import lxml.etree as etree
 import xml.etree.ElementTree as ET
 from collections.abc import MutableMapping
 from typing_extensions import Self
@@ -34,6 +35,9 @@ __all__ = ["SdkJSONEncoder", "Model", "rest_field", "rest_discriminator"]
 TZ_UTC = timezone.utc
 _T = typing.TypeVar("_T")
 
+_XML_PARSER = etree.XMLParser(resolve_entities=False)
+"""Avoid XML entity resolution to prevent XXE attacks. (https://github.com/microsoft/typespec/issues/8083)"""
+
 
 def _timedelta_as_isostr(td: timedelta) -> str:
     """Converts a datetime.timedelta object into an ISO 8601 formatted string, e.g. 'P4DT12H30M05S'
@@ -959,7 +963,7 @@ def _failsafe_deserialize(
 
 def _failsafe_deserialize_xml(
   deserializer: typing.Any,
-  response: HttpResponse, 
+  response: HttpResponse,
 ) -> typing.Any:
   try:
       return _deserialize_xml(deserializer, response.text())
@@ -1208,7 +1212,7 @@ def _deserialize_xml(
     deserializer: typing.Any,
     value: str,
 ) -> typing.Any:
-    element = ET.fromstring(value) # nosec
+    element = etree.fromstring(value, parser=_XML_PARSER)
     return _deserialize(deserializer, element)
 
 

packages/http-client-python/generator/pygen/codegen/templates/serialization.py.jinja2

@@ -32,6 +32,7 @@ try:
     from urllib import quote  # type: ignore
 except ImportError:
     from urllib.parse import quote
+import lxml.etree as etree
 import xml.etree.ElementTree as ET
 
 import isodate  # type: ignore
@@ -44,6 +45,9 @@ _BOM = codecs.BOM_UTF8.decode(encoding="utf-8")
 
 JSON = MutableMapping[str, Any]
 
+_XML_PARSER = etree.XMLParser(resolve_entities=False)
+"""Avoid XML entity resolution to prevent XXE attacks. (https://github.com/microsoft/typespec/issues/8083)"""
+
 
 class RawDeserializer:
 
@@ -98,7 +102,7 @@ class RawDeserializer:
                 except NameError:
                     pass
 
-                return ET.fromstring(data_as_str)  # nosec
+                return etree.fromstring(data_as_str, parser=_XML_PARSER)
             except ET.ParseError as err:
                 # It might be because the server has an issue, and returned JSON with
                 # content-type XML....

packages/http-client-python/generator/setup.py

@@ -53,5 +53,6 @@
         "PyYAML==6.0.1",
         "tomli==2.0.1",
         "setuptools==75.3.2",
+        "lxml>=6.0,<7.0"
     ],
 )