--- Begin Message ---
Package: release.debian.org
Severity: normal
Tags: trixie
X-Debbugs-Cc: [email protected], [email protected]
Control: affects -1 + src:lxml-html-clean
User: [email protected]
Usertags: pu
- CVE-2026-28348: CSS @import Filter Bypass via Unicode Escapes
- CVE-2026-28350: <base> tag injection through default Cleaner
configuration
The only code changes in the new upstream releases are the CVE fixes
in clean.py, everything else are test/CI/documentation changes
(including testcases for the CVEs).
diffstat for lxml-html-clean-0.4.2 lxml-html-clean-0.4.4
.github/workflows/main.yml | 4
CHANGES.rst | 25 +++++
README.md | 2
debian/changelog | 25 +++++
debian/control | 2
docs/usage.rst | 5 +
lxml_html_clean/clean.py | 27 ++++++
setup.cfg | 5 -
tests/test_clean.py | 198 ++++++++++++++++++++++++++++++++++++++++++++-
tests/test_clean.txt | 2
tox.ini | 2
11 files changed, 284 insertions(+), 13 deletions(-)
diff -Nru lxml-html-clean-0.4.2/CHANGES.rst lxml-html-clean-0.4.4/CHANGES.rst
--- lxml-html-clean-0.4.2/CHANGES.rst 2025-04-09 14:14:25.000000000 +0300
+++ lxml-html-clean-0.4.4/CHANGES.rst 2026-02-27 11:32:37.000000000 +0200
@@ -6,6 +6,31 @@
Unreleased
==========
+0.4.4 (2026-02-26)
+==================
+
+Bugs fixed
+----------
+
+* Fixed a bug where Unicode escapes in CSS were not properly decoded
+ before security checks. This prevents attackers from bypassing filters
+ using escape sequences.
+* Fixed a security issue where ``<base>`` tags could be used for URL
+ hijacking attacks. The ``<base>`` tag is now automatically removed
+ whenever the ``<head>`` tag is removed (via ``page_structure=True``
+ or manual configuration), as ``<base>`` must be inside ``<head>``
+ according to HTML specifications.
+
+0.4.3 (2025-10-02)
+==================
+
+Maintenance
+-----------
+
+* Tests updated to work correctly with new lxml and libxml2 releases.
+* Python 3.6 and 3.7 are no longer tested.
+* Improved documentation about CSS removal behavior.
+
0.4.2 (2025-04-09)
==================
diff -Nru lxml-html-clean-0.4.2/debian/changelog
lxml-html-clean-0.4.4/debian/changelog
--- lxml-html-clean-0.4.2/debian/changelog 2025-04-23 08:47:04.000000000
+0300
+++ lxml-html-clean-0.4.4/debian/changelog 2026-06-18 20:48:59.000000000
+0300
@@ -1,3 +1,28 @@
+lxml-html-clean (0.4.4-1~deb13u1) trixie; urgency=medium
+
+ * Non-maintainer upload.
+ * Rebuild for trixie.
+ - CVE-2026-28348: CSS @import Filter Bypass via Unicode Escapes
+ - CVE-2026-28350: <base> tag injection through default Cleaner
+ configuration
+
+
+ -- Adrian Bunk <[email protected]> Thu, 18 Jun 2026 20:48:59 +0300
+
+lxml-html-clean (0.4.4-1) unstable; urgency=medium
+
+ * New upstream version.
+ * Bump standards version.
+
+ -- Matthias Klose <[email protected]> Fri, 06 Mar 2026 08:37:31 +0100
+
+lxml-html-clean (0.4.3-1) unstable; urgency=medium
+
+ * New upstream version. Closes: #1114193.
+ * Bump standards version.
+
+ -- Matthias Klose <[email protected]> Sun, 05 Oct 2025 11:15:18 +0200
+
lxml-html-clean (0.4.2-1) unstable; urgency=medium
* New upstream version.
diff -Nru lxml-html-clean-0.4.2/debian/control
lxml-html-clean-0.4.4/debian/control
--- lxml-html-clean-0.4.2/debian/control 2024-08-02 03:41:19.000000000
+0300
+++ lxml-html-clean-0.4.4/debian/control 2026-03-06 09:37:31.000000000
+0200
@@ -9,7 +9,7 @@
python3-setuptools,
python3-all,
python3-lxml,
-Standards-Version: 4.7.0
+Standards-Version: 4.7.3
Homepage: https://github.com/fedora-python/lxml_html_clean
Package: python3-lxml-html-clean
diff -Nru lxml-html-clean-0.4.2/docs/usage.rst
lxml-html-clean-0.4.4/docs/usage.rst
--- lxml-html-clean-0.4.2/docs/usage.rst 2025-04-09 14:14:25.000000000
+0300
+++ lxml-html-clean-0.4.4/docs/usage.rst 2026-02-27 11:32:37.000000000
+0200
@@ -109,6 +109,11 @@
</body>
</html>
+To control the removal of CSS styles, set the ``style`` and/or ``inline_style``
+keyword arguments to ``True`` when creating a ``Cleaner`` instance.
+If neither option is enabled, only ``@import`` rules are automatically removed
+from CSS content.
+
You can also whitelist some otherwise dangerous content with
``Cleaner(host_whitelist=['www.youtube.com'])``, which would allow
embedded media from YouTube, while still filtering out embedded media
diff -Nru lxml-html-clean-0.4.2/.github/workflows/main.yml
lxml-html-clean-0.4.4/.github/workflows/main.yml
--- lxml-html-clean-0.4.2/.github/workflows/main.yml 2025-04-09
14:14:25.000000000 +0300
+++ lxml-html-clean-0.4.4/.github/workflows/main.yml 2026-02-27
11:32:37.000000000 +0200
@@ -19,17 +19,15 @@
uses: fedora-python/tox-github-action@main
with:
tox_env: ${{ matrix.tox_env }}
- dnf_install: gcc libxml2-devel libxslt-devel
strategy:
matrix:
tox_env:
- - py36
- - py38
- py39
- py310
- py311
- py312
- py313
+ - py314
- mypy
# Use GitHub's Linux Docker host
diff -Nru lxml-html-clean-0.4.2/lxml_html_clean/clean.py
lxml-html-clean-0.4.4/lxml_html_clean/clean.py
--- lxml-html-clean-0.4.2/lxml_html_clean/clean.py 2025-04-09
14:14:25.000000000 +0300
+++ lxml-html-clean-0.4.4/lxml_html_clean/clean.py 2026-02-27
11:32:37.000000000 +0200
@@ -422,6 +422,12 @@
if self.annoying_tags:
remove_tags.update(('blink', 'marquee'))
+ # Remove <base> tags whenever <head> is being removed.
+ # According to HTML spec, <base> must be in <head>, but browsers
+ # may interpret it even when misplaced, allowing URL hijacking attacks.
+ if 'head' in kill_tags or 'head' in remove_tags:
+ kill_tags.add('base')
+
_remove = deque()
_kill = deque()
for el in doc.iter():
@@ -578,6 +584,26 @@
_comments_re = re.compile(r'/\*.*?\*/', re.S)
_find_comments = _comments_re.finditer
_substitute_comments = _comments_re.sub
+ _css_unicode_escape_re = re.compile(r'\\([0-9a-fA-F]{1,6})\s?')
+
+ def _decode_css_unicode_escapes(self, style):
+ """
+ Decode CSS Unicode escape sequences like \\69 or \\000069 to their
+ actual character values. This prevents bypassing security checks
+ using CSS escape sequences.
+
+ CSS escape syntax: backslash followed by 1-6 hex digits,
+ optionally followed by a whitespace character.
+ """
+ def replace_escape(match):
+ hex_value = match.group(1)
+ try:
+ return chr(int(hex_value, 16))
+ except (ValueError, OverflowError):
+ # Invalid unicode codepoint, keep original
+ return match.group(0)
+
+ return self._css_unicode_escape_re.sub(replace_escape, style)
def _has_sneaky_javascript(self, style):
"""
@@ -591,6 +617,7 @@
more sneaky attempts.
"""
style = self._substitute_comments('', style)
+ style = self._decode_css_unicode_escapes(style)
style = style.replace('\\', '')
style = _substitute_whitespace('', style)
style = style.lower()
diff -Nru lxml-html-clean-0.4.2/README.md lxml-html-clean-0.4.4/README.md
--- lxml-html-clean-0.4.2/README.md 2025-04-09 14:14:25.000000000 +0300
+++ lxml-html-clean-0.4.4/README.md 2026-02-27 11:32:37.000000000 +0200
@@ -4,7 +4,7 @@
This project was initially a part of [lxml](https://github.com/lxml/lxml).
Because HTML cleaner is designed as blocklist-based, many reports about
possible security vulnerabilities were filed for lxml and that make the project
problematic for security-sensitive environments. Therefore we decided to
extract the problematic part to a separate project.
-**Important**: the HTML Cleaner in ``lxml_html_clean`` is **not** considered
appropriate **for security sensitive environments**. See e.g.
[bleach](https://pypi.org/project/bleach/) for an alternative.
+**Important**: the HTML Cleaner in ``lxml_html_clean`` is **not** considered
appropriate **for security sensitive environments**. See e.g.
[nh3](https://pypi.org/project/nh3/) for an alternative.
This project uses functions from Python's `urllib.parse` for URL parsing which
**do not validate inputs**. For more information on potential security risks,
refer to the [URL parsing
security](https://docs.python.org/3/library/urllib.parse.html#url-parsing-security)
documentation. A maliciously crafted URL could potentially bypass the allowed
hosts check in `Cleaner`.
diff -Nru lxml-html-clean-0.4.2/setup.cfg lxml-html-clean-0.4.4/setup.cfg
--- lxml-html-clean-0.4.2/setup.cfg 2025-04-09 14:14:25.000000000 +0300
+++ lxml-html-clean-0.4.4/setup.cfg 2026-02-27 11:32:37.000000000 +0200
@@ -1,6 +1,6 @@
[metadata]
name = lxml_html_clean
-version = 0.4.2
+version = 0.4.4
description = HTML cleaner from lxml project
long_description = file:README.md
long_description_content_type = text/markdown
@@ -13,14 +13,13 @@
license_files = LICENSE.txt
classifiers =
Programming Language :: Python :: 3
- Programming Language :: Python :: 3.6
- Programming Language :: Python :: 3.7
Programming Language :: Python :: 3.8
Programming Language :: Python :: 3.9
Programming Language :: Python :: 3.10
Programming Language :: Python :: 3.11
Programming Language :: Python :: 3.12
Programming Language :: Python :: 3.13
+ Programming Language :: Python :: 3.14
[options]
packages =
diff -Nru lxml-html-clean-0.4.2/tests/test_clean.py
lxml-html-clean-0.4.4/tests/test_clean.py
--- lxml-html-clean-0.4.2/tests/test_clean.py 2025-04-09 14:14:25.000000000
+0300
+++ lxml-html-clean-0.4.4/tests/test_clean.py 2026-02-27 11:32:37.000000000
+0200
@@ -331,20 +331,20 @@
def test_host_whitelist_valid(self):
# Frame with valid hostname in src is allowed.
- html = '<div><iframe src="https://example.com/page";></div>'
+ html = '<div><iframe src="https://example.com/page";></iframe></div>'
expected = '<div><iframe
src="https://example.com/page";></iframe></div>'
cleaner = Cleaner(frames=False, host_whitelist=["example.com"])
self.assertEqual(expected, cleaner.clean_html(html))
def test_host_whitelist_invalid(self):
- html = '<div><iframe src="https://evil.com/page";></div>'
+ html = '<div><iframe src="https://evil.com/page";></iframe></div>'
expected = '<div></div>'
cleaner = Cleaner(frames=False, host_whitelist=["example.com"])
self.assertEqual(expected, cleaner.clean_html(html))
def test_host_whitelist_sneaky_userinfo(self):
# Regression test: Don't be fooled by hostname and colon in userinfo.
- html = '<div><iframe src="https://example.com:@evil.com/page";></div>'
+ html = '<div><iframe
src="https://example.com:@evil.com/page";></iframe></div>'
expected = '<div></div>'
cleaner = Cleaner(frames=False, host_whitelist=["example.com"])
self.assertEqual(expected, cleaner.clean_html(html))
@@ -393,3 +393,195 @@
self.assertEqual(len(w), 0)
self.assertNotIn("google.com", result)
self.assertNotIn("example.com", result)
+
+ def test_base_tag_removed_with_page_structure(self):
+ # Test that <base> tags are removed when page_structure=True (default)
+ # This prevents URL hijacking attacks where <base> redirects all
relative URLs
+
+ test_cases = [
+ # <base> in proper location (inside <head>)
+ '<html><head><base href="http://evil.com/";></head><body><a
href="page.html">link</a></body></html>',
+ # <base> outside <head>
+ '<div><base href="http://evil.com/";><a
href="page.html">link</a></div>',
+ # Multiple <base> tags
+ '<base href="http://evil.com/";><div><base
href="http://evil2.com/";></div>',
+ # <base> with target attribute
+ '<base target="_blank"><div>content</div>',
+ # <base> at various positions
+ '<html><base href="http://evil.com/";><body>test</body></html>',
+ ]
+
+ for html in test_cases:
+ with self.subTest(html=html):
+ cleaned = clean_html(html)
+ # Verify <base> tag is completely removed
+ self.assertNotIn('base', cleaned.lower())
+ self.assertNotIn('evil.com', cleaned)
+ self.assertNotIn('evil2.com', cleaned)
+
+ def test_base_tag_kept_when_page_structure_false(self):
+ # When page_structure=False and head is not removed, <base> should be
kept
+ cleaner = Cleaner(page_structure=False)
+ html = '<html><head><base
href="http://example.com/";></head><body>test</body></html>'
+ cleaned = cleaner.clean_html(html)
+ self.assertIn('<base href="http://example.com/";>', cleaned)
+
+ def test_base_tag_removed_when_head_in_remove_tags(self):
+ # Even with page_structure=False, <base> should be removed if head is
manually removed
+ cleaner = Cleaner(page_structure=False, remove_tags=['head'])
+ html = '<html><head><base
href="http://evil.com/";></head><body>test</body></html>'
+ cleaned = cleaner.clean_html(html)
+ self.assertNotIn('base', cleaned.lower())
+ self.assertNotIn('evil.com', cleaned)
+
+ def test_base_tag_removed_when_head_in_kill_tags(self):
+ # Even with page_structure=False, <base> should be removed if head is
in kill_tags
+ cleaner = Cleaner(page_structure=False, kill_tags=['head'])
+ html = '<html><head><base
href="http://evil.com/";></head><body>test</body></html>'
+ cleaned = cleaner.clean_html(html)
+ self.assertNotIn('base', cleaned.lower())
+ self.assertNotIn('evil.com', cleaned)
+
+ def test_unicode_escape_in_style(self):
+ # Test that CSS Unicode escapes are properly decoded before security
checks
+ # This prevents attackers from bypassing filters using escape sequences
+ # CSS escape syntax: \HHHHHH where H is a hex digit (1-6 digits)
+
+ # Test inline style attributes (requires safe_attrs_only=False)
+ cleaner = Cleaner(safe_attrs_only=False)
+ inline_style_cases = [
+ # \6a\61\76\61\73\63\72\69\70\74 = "javascript"
+ ('<div style="background:
url(\\6a\\61\\76\\61\\73\\63\\72\\69\\70\\74:alert(1))">test</div>',
'<div>test</div>'),
+ # \69 = 'i', so \69mport = "import"
+ ('<div style="@\\69mport url(evil.css)">test</div>',
'<div>test</div>'),
+ # \69 with space after = 'i', space consumed as part of escape
+ ('<div style="@\\69 mport url(evil.css)">test</div>',
'<div>test</div>'),
+ # \65\78\70\72\65\73\73\69\6f\6e = "expression"
+ ('<div
style="\\65\\78\\70\\72\\65\\73\\73\\69\\6f\\6e(alert(1))">test</div>',
'<div>test</div>'),
+ ]
+
+ for html, expected in inline_style_cases:
+ with self.subTest(html=html):
+ cleaned = cleaner.clean_html(html)
+ self.assertEqual(expected, cleaned)
+
+ # Test <style> tag content (uses default clean_html)
+ style_tag_cases = [
+ # Unicode-escaped "javascript:" in url()
+
'<style>url(\\6a\\61\\76\\61\\73\\63\\72\\69\\70\\74:alert(1))</style>',
+ # Unicode-escaped "javascript:" without url()
+ '<style>\\6a\\61\\76\\61\\73\\63\\72\\69\\70\\74:alert(1)</style>',
+ # Unicode-escaped "expression"
+
'<style>\\65\\78\\70\\72\\65\\73\\73\\69\\6f\\6e(alert(1))</style>',
+ # Unicode-escaped @import with 'i'
+ '<style>@\\69mport url(evil.css)</style>',
+ # Unicode-escaped "data:" scheme
+
'<style>url(\\64\\61\\74\\61:image/svg+xml;base64,PHN2ZyBvbmxvYWQ9YWxlcnQoMSk+)</style>',
+ # Space after escape is consumed: \69 mport = "import"
+ '<style>@\\69 mport url(evil.css)</style>',
+ # 6-digit escape: \000069 = 'i'
+ '<style>@\\000069mport url(evil.css)</style>',
+ # 6-digit escape with space
+ '<style>@\\000069 mport url(evil.css)</style>',
+ ]
+
+ for html in style_tag_cases:
+ with self.subTest(html=html):
+ cleaned = clean_html(html)
+ self.assertEqual('<div><style>/* deleted */</style></div>',
cleaned)
+
+ def test_unicode_escape_mixed_with_comments(self):
+ # Unicode escapes mixed with CSS comments should still be caught
+ test_cases = [
+ # \69 = 'i' with comment before
+ '<style>@/*comment*/\\69mport url(evil.css)</style>',
+ # \69 = 'i' with comment after
+ '<style>@\\69mport/*comment*/ url(evil.css)</style>',
+ # Multiple escapes with comments
+
'<style>\\65\\78/*comment*/\\70\\72\\65\\73\\73\\69\\6f\\6e(alert(1))</style>',
+ ]
+
+ for html in test_cases:
+ with self.subTest(html=html):
+ cleaned = clean_html(html)
+ self.assertEqual('<div><style>/* deleted */</style></div>',
cleaned)
+
+ def test_unicode_escape_case_insensitive(self):
+ # CSS hex escapes should work with both uppercase and lowercase hex
digits
+ # \69 = 'i', \6D = 'm', etc.
+ test_cases = [
+ # @import with uppercase hex digits: \69\6D\70\6F\72\74
+ '<style>@\\69\\6D\\70\\6F\\72\\74 url(evil.css)</style>',
+ # @import with some uppercase
+ '<style>@\\69\\6D\\70\\6f\\72\\74 url(evil.css)</style>',
+ ]
+
+ for html in test_cases:
+ with self.subTest(html=html):
+ cleaned = clean_html(html)
+ self.assertEqual('<div><style>/* deleted */</style></div>',
cleaned)
+
+ def test_unicode_escape_various_schemes(self):
+ # Test Unicode escapes for various malicious schemes
+ test_cases = [
+ # \76\62\73\63\72\69\70\74 = "vbscript"
+ '<style>url(\\76\\62\\73\\63\\72\\69\\70\\74:alert(1))</style>',
+ # \6a\73\63\72\69\70\74 = "jscript"
+ '<style>url(\\6a\\73\\63\\72\\69\\70\\74:alert(1))</style>',
+ # \6c\69\76\65\73\63\72\69\70\74 = "livescript"
+
'<style>url(\\6c\\69\\76\\65\\73\\63\\72\\69\\70\\74:alert(1))</style>',
+ # \6d\6f\63\68\61 = "mocha"
+ '<style>url(\\6d\\6f\\63\\68\\61:alert(1))</style>',
+ ]
+
+ for html in test_cases:
+ with self.subTest(html=html):
+ cleaned = clean_html(html)
+ self.assertEqual('<div><style>/* deleted */</style></div>',
cleaned)
+
+ def test_unicode_escape_with_whitespace_variations(self):
+ # Test different whitespace characters after Unicode escapes
+ cleaner = Cleaner(safe_attrs_only=False)
+ test_cases = [
+ # Tab after escape
+ ('<div style="@\\69\tmport url(evil.css)">test</div>',
'<div>test</div>'),
+ # Newline after escape (note: actual newline, not \n)
+ ('<div style="@\\69\nmport url(evil.css)">test</div>',
'<div>test</div>'),
+ # Form feed after escape
+ ('<div style="@\\69\fmport url(evil.css)">test</div>',
'<div>test</div>'),
+ ]
+
+ for html, expected in test_cases:
+ with self.subTest(html=html):
+ cleaned = cleaner.clean_html(html)
+ self.assertEqual(expected, cleaned)
+
+ def test_backslash_removal_after_unicode_decode(self):
+ # After decoding Unicode escapes, remaining backslashes are removed
+ # This ensures double-obfuscation (unicode + backslashes) is caught
+ test_cases = [
+ # Step 1: \69 → 'i', Step 2: remove \, Result: @import
+ '<style>@\\69\\m\\p\\o\\r\\t url(evil.css)</style>',
+ # Multiple unicode escapes with backslashes mixed in
+ '<style>@\\69\\6d\\p\\6f\\r\\t url(evil.css)</style>',
+ ]
+
+ for html in test_cases:
+ with self.subTest(html=html):
+ cleaned = clean_html(html)
+ self.assertEqual('<div><style>/* deleted */</style></div>',
cleaned)
+
+ def test_backslash_obfuscation_without_unicode(self):
+ # Test that patterns using ONLY backslash obfuscation (no unicode) are
caught
+ # Step 1: No unicode escapes, Step 2: remove \, Result: malicious
pattern
+ test_cases = [
+ # @\i\m\p\o\r\t → @import (caught by '@import' check)
+ '<style>@\\i\\m\\p\\o\\r\\t url(evil.css)</style>',
+ # Can also test combinations that create javascript schemes
+ '<style>@\\import url(evil.css)</style>',
+ ]
+
+ for html in test_cases:
+ with self.subTest(html=html):
+ cleaned = clean_html(html)
+ self.assertEqual('<div><style>/* deleted */</style></div>',
cleaned)
diff -Nru lxml-html-clean-0.4.2/tests/test_clean.txt
lxml-html-clean-0.4.4/tests/test_clean.txt
--- lxml-html-clean-0.4.2/tests/test_clean.txt 2025-04-09 14:14:25.000000000
+0300
+++ lxml-html-clean-0.4.4/tests/test_clean.txt 2026-02-27 11:32:37.000000000
+0200
@@ -84,7 +84,7 @@
<body onload="evil_function()">
<!-- I am interpreted for EVIL! -->
<a href="javascript:evil_function()">a link</a>
- <a href="javascrip%20t%20:evil_function()">a control char link</a>
+ <a href="j%01a%02v%03a%04s%05c%06r%07i%0Ep%20t%20:evil_function()">a
control char link</a>
<a
href="data:text/html;base64,PHNjcmlwdD5hbGVydCgidGVzdCIpOzwvc2NyaXB0Pg==">data</a>
<a href="#" onclick="evil_function()">another link</a>
<p onclick="evil_function()">a paragraph</p>
diff -Nru lxml-html-clean-0.4.2/tox.ini lxml-html-clean-0.4.4/tox.ini
--- lxml-html-clean-0.4.2/tox.ini 2025-04-09 14:14:25.000000000 +0300
+++ lxml-html-clean-0.4.4/tox.ini 2026-02-27 11:32:37.000000000 +0200
@@ -1,5 +1,5 @@
[tox]
-envlist = py36,py38,py39,py310,py311,py312,py313,mypy
+envlist = py39,py310,py311,py312,py313,py314,mypy
skipsdist = True
[testenv]
--- End Message ---
