More robust HTML sanitizing in reports (#12997)

src/backend/InvenTree/report/templatetags/report.py

@@ -18,10 +18,12 @@ from django.core.files.storage import default_storage
 from django.db.models import Model
 from django.db.models.query import QuerySet
 from django.utils import translation
+from django.utils.html import conditional_escape, format_html
 from django.utils.safestring import SafeString, mark_safe
 from django.utils.translation import gettext_lazy as _
 
 import lxml.html
+import nh3
 from babel import Locale
 from babel.core import UnknownLocaleError
 from babel.dates import format_date as babel_format_date
@@ -766,9 +768,9 @@ def internal_link(link, text) -> str:
     # If the base URL is not set, just return the text
     if not url:
         logger.warning('Failed to construct absolute URL for internal link')
-        return text
+        return conditional_escape(text)
 
-    return mark_safe(f'<a href="{url}">{text}</a>')
+    return format_html('<a href="{}">{}</a>', url, text)
 
 
 def make_decimal(value: Any) -> Any:
@@ -1116,6 +1118,7 @@ def render_html_text(text: str, **kwargs):
         bold: Boolean, whether bold (or not)
         italic: Boolean, whether italic (or not)
         heading: str, heading level e.g. 'h3'
+        sanitize: Boolean, whether to sanitize the provided text (default = True)
     """
     tags = []
 
@@ -1128,6 +1131,12 @@ def render_html_text(text: str, **kwargs):
     if heading := kwargs.get('heading', ''):
         tags.append(heading)
 
+    text = str(text)
+
+    if kwargs.get('sanitize', True):
+        # Remove any unsafe HTML content (e.g. script tags) from the provided text
+        text = nh3.clean(text)
+
     output = ''.join([f'<{tag}>' for tag in tags])
     output += text
     output += ''.join([f'</{tag}>' for tag in tags])

src/backend/InvenTree/report/test_tags.py

@@ -326,6 +326,17 @@ class ReportTagTest(PartImageTestMixin, InvenTreeTestCase):
         link = report_tags.internal_link(None, None)
         self.assertEqual(link, 'None')
 
+        # Unsafe text content is escaped
+        payload = '"><script>alert(1)</script>'
+        escaped = '&quot;&gt;&lt;script&gt;alert(1)&lt;/script&gt;'
+
+        link = report_tags.internal_link(obj.get_absolute_url(), payload)
+        self.assertIn(f'>{escaped}</a>', link)
+        self.assertNotIn('<script>', link)
+
+        # Unsafe text is also escaped when no link can be constructed
+        self.assertEqual(report_tags.internal_link(None, payload), escaped)
+
     def test_logo_image(self):
         """Unit tests for the 'logo_image' tag."""
         # In debug mode, a web-accessible URL to the logo is returned
@@ -965,6 +976,24 @@ class ReportTagTest(PartImageTestMixin, InvenTreeTestCase):
             '<h1>hello world</h1>',
         )
 
+        # Unsafe HTML content is sanitized by default
+        payload = (
+            '<b>hello</b><script>alert(1)</script><img src="x" onerror="alert(1)">'
+        )
+
+        self.assertEqual(
+            report_tags.render_html_text(payload, bold=True),
+            '<strong><b>hello</b><img src="x"></strong>',
+        )
+
+        # Encoded HTML entities must not be decoded into live tags
+        self.assertEqual(
+            report_tags.render_html_text('&lt;script&gt;'), '&lt;script&gt;'
+        )
+
+        # Sanitization can be explicitly disabled
+        self.assertEqual(report_tags.render_html_text(payload, sanitize=False), payload)
+
     def test_format_date(self):
         """Test the format_date template tag."""
         dt = timezone.datetime(year=2024, month=3, day=13)