
A previous rework of the directory traversal mitigation example in I3f8d3760daceb9e62396ae21b0d915ae07eff303 was not correctly cleaned up, and left some unintended startswith method invocations behind. Get rid of those, and also correct a wrong parameter name in the main function while we're at it, as well as fixing some incorrect indentation. Change-Id: Ie5347f3b6cc8e689440db0aaf552d52ad37c231c Closes-Bug: #1928544
4.3 KiB
Restrict path access to prevent path traversal
Often we will refer to a file on disk or other resource using a path. A path traversal attack is when an attacker supplies input that gets used with our path to access a file on the file system that we did not intend. The input usually attempts to break out of the application's working directory and access a file elsewhere on the file system. This category of attack can be mitigated by restricting the scope of file system access and reducing attack surface by using a restricted file permission profile.
Incorrect
A typical remote vector for path traversal in web applications might involve serving or storing files on the file system. Consider the following example:
import os
from flask import Flask, redirect, request, send_file
= Flask(__name__)
app
@app.route('/')
def cat_picture():
= request.args.get('image_name')
image_name if not image_name:
return 404
return send_file(os.path.join(os.getcwd(), image_name))
if __name__ == '__main__':
=True) app.run(debug
As the attacker controls the input that is used directly in constructing a path they are able to access any file on the system. For example consider what happens if an attacker makes a request like:
curl http://example.com/?image_name=../../../../../../../../etc/passwd
Path traversal flaws also can happen when unpacking a compressed archive of files. An example of where this has happened within OpenStack is OSSA-2011-001. In this case a tar file from an untrusted source could be unpacked to overwrite files on the host operating system.
import tarfile
def untar_image(path, filename):
= tarfile.open(filename, 'r|gz')
tar_file
tar_file.extract_all(path)= tar_file.get_names()[0]
image_file
tar_file.close()return os.path.join(path, image_file)
Correct
The following example demonstrates how you can use python code to restrict access to files within a specific directory. This can be used as a mechanism to defeat path traversal.
import os
import sys
def is_safe_path(basedir, path, follow_symlinks=True):
# resolves symbolic links
if follow_symlinks:
= os.path.realpath(path)
matchpath else:
= os.path.abspath(path)
matchpath return basedir == os.path.commonpath((basedir, matchpath))
def main(args):
for arg in args:
if is_safe_path(os.getcwd(), arg):
print("safe: {}".format(arg))
else:
print("unsafe: {}".format(arg))
if __name__ == "__main__":
1:]) main(sys.argv[
Another approach to restricting file system access to maintain an indirect mapping between a unique identifier and a file path that exists on the operating system. This prevents users supplying malicious input to access unintended files.
= {
localfiles "01" : "/var/www/img/001.png",
"02" : "/var/www/img/002.png",
"03" : "/var/www/img/003.png",
}
# Will raise an error if an invalid key is used.
def get_file(file_id):
return open(localfiles[file_id])
Consequences
Not validating file paths allows the attacker to read or write to any file that the application has access to. This can lead to information leakage and can be used to pivot to other more serious attacks like remote code execution.