As a part of security assessment of cloud based apps/ infrastructure we always face a challenge in scanning the servers in the cloud. Few of them are:
Obtaining/ managing credentials always an headache
Not ideal for cloud solutions
Requires target machines to be always online
The limitations of the scanners:
Traditional infrastructure scanners such as Nessus are of not much use
Sometimes the scanners does not report vulnerabilities correctly due to many issues such as machines frequently go down while scan is in process, some firewall issues etc
We need a solution which is rather than we scanning the target servers, it resides on the server and keep doing the scanning and sends the report back to the organization periodically. And here comes the concept of 'Agent based clod scanning'.
The benefits are:
Rather than targeting the remote servers in the traditional approach, the agents installed on the servers keep on doing the scans and sends the periodic reports
Can run offline and syncs when becomes online
It helps reducing the network congestion
Enumerates Bluetooth, USB devices and mounted shares
Detects malwares and suspicious processes
Nessus recently launched such tool called, Nessus Agents which fulfills above conditions.
Among many authentication modes for accessing resources over cloud, such as traditional authentication such as credentials, or muti-factor authentication, such as hardware tokens; the biggest issue is that they can be stolen, or mimicked. The traditional solutions available in market are mimicable and not fool proof, the hardware tokens, passwords etc. are easy to compromise. Also, the traditional approach towards the authentication process- first authentication via user credentials then use of any other mode of authentication such as hardware token- increases the attack surface.
How about reversing the above approach- first people who can prove who they are (Biometric) only can access the Login page. This will decrease the risk significantly as the login page will be available to a very few set of people rather than whole bunch.
So the steps are:
Biometric authentication- adding ‘what- you- are- factor’
Raises the security bar to the highest level
Challenging the traditional way of implementing multi-factor authentication:
1. First biometric authentication
2. Followed by, any traditional mode of authentication- passwords, tokens etc.
It reduces the probability of attack
So, how does PCI-DSS affects our web application security testing or what to make sure the application is compliant with PCI, while doing the security testing.
Here are few requirements which needs to be taken case while testing a web application which handles financial data such as credit card information. As the PCI guidelines itself maintains that the application must be tested on regular basis in "Requirement 11: Regularly test security systems and processes."
But what should really we look for? The requirement 11 is tied back to other requirements Requirement 6.:
11.2.1 Perform quarterly internal vulnerability scans and rescans as needed, until all “high-risk” vulnerabilities (as identified in Requirement 6.1) are resolved. Scans must be performed by qualified personnel.
11.2.3 Perform internal and external scans, and rescans as needed, after any significant change. Scans must be performed by qualified personnel.
11.3 Implement a methodology for penetration testing that includes the following:
11.3.1 Perform external penetration testing at least annually and after any significant infrastructure or application upgrade or modification (such as an operating system upgrade, a sub-network added to the environment, or a web server added to the environment).
11.3.2 Perform internal penetration testing at least annually and after any significant infrastructure or application upgrade or modification (such as an operating system upgrade, a sub-network added to the environment, or a web server added to the environment).
Now let's check Requirement 6.1 says:
6.1 Establish a process to identify security vulnerabilities, using reputable outside sources for security vulnerability information, and assign a risk ranking (for example, as “high,” “medium,” or “low”) to newly discovered security vulnerabilities.
6.2 Ensure that all system components and software are protected from known vulnerabilities by installing applicable vendor-supplied security patches. Install critical security patches within one month of release.
6.3 Develop internal and external software applications (including web-based administrative access to applications) securely, as follows.
And what to test for:
6.5 Address common coding vulnerabilities in software-development processes as follows:
6.5.1 Injection flaws, particularly SQL injection. Also consider OS Command Injection, LDAP and XPath injection flaws as well as other injection flaws.
6.5.2 Buffer overflows
6.5.3 Insecure cryptographic storage
6.5.4 Insecure communications
6.5.5 Improper error handling
6.5.6 All “high risk” vulnerabilities identified in the vulnerability identification process
6.5.7 Cross-site scripting (XSS)
6.5.8 Improper access control
6.5.9 Cross-site request forgery (CSRF)
6.5.10 Broken authentication and session management
Basically we need to test the web applucation for OWASP Top 10 vulnerabilities and addtionally we need to scan the servers and infrastructure for possible open ports, patches missing any other misconfigurations.
So basically, focus on Requirement section 6 while doing the testing.
There could be some instances where in you need to scan your Android devices with scanners such as Nessus etc to look for insecure/ unnecessary ports, services and misconfigurations.
There are two types of scanning- unauthenticated scan and authenticated scan. Unauthenticated scans are preatty simple, just provide the IP of the target to be scanned, but in case of an authenticated scan which is more comprehensive, you need to have some valid account created on the target device. So, how to run an authenticated scan on Android device? We don't have any IS level account on it.
One way to accomplish this is to create an ssh server on the device. Once the server is installed, it is very basic to run ssh commands remotely such as we do using Putty.
The steps are following:
1. Go ahead and download, install an ssh server. ssh servers such as SSHDroid, SSHelper etc can be installed. They can be installed via Google Play.
One is here:
2. Create an account on the server.
3. Provide the account credentials to Nessus for an authenticated scan.
It's a very trivial mistake a lot of developers do. They don't pay much attention to this simple looking misconfigurations which can be a big risk to apps. Generally they put much attention on other best practices such as intents, permissions etc, but tend to ignore the debuggable and allowbackup settings. I came across various such mistakes.
The app is debuggable, which means we can attach a debugger to the process and step through every single instruction and even execute arbitrary code in the context of app process.
Similarly, allowbackup is used to determine if to allow the application to participate in the backup and restore infrastructure. This leads to potential data leakage.
If not required, harden the configuration file by setting the values= “False”
This is a tricky situation as you can't see the page source to look for the tags such as embed etc. All you see is the entire page loads as Flash content and plugins like InspectElement will not work.
I came across the similar scenario.
So, this is how we can sort out this issue. Every browser has flash plugins to display flash contents in the browser itself. So, if we disable the plugin the browser won't be able to display it within and will attempt to download it on harddisk, which can be used for security analysis.
Here how to do in case of Chrome. Just disable the plugin from the browser: