Application Discovery & Enumeration
Initial Enumeration
Let's assume our client provided us with the following scope:
We can start with an Nmap scan of common web ports. I'll typically do an initial scan with ports 80,443,8000,8080,8180,8888,10000 and then run either EyeWitness or Aquatone (or both depending on the results of the first) against this initial scan.
While reviewing the screenshot report of the most common ports, I may run a more thorough Nmap scan against the top 10,000 ports or all TCP ports, depending on the size of the scope. Since enumeration is an iterative process, we will run a web screenshotting tool against any subsequent Nmap scans we perform to ensure maximum coverage.
On a non-evasive full scope penetration test, I will usually run a Nessus scan too to give the client the most bang for their buck, but we must be able to perform assessments without relying on scanning tools. Even though most assessments are time-limited (and often not scoped appropriately for the size of the environment), we can provide our clients maximum value by establishing a repeatable and thorough enumeration methodology that can be applied to all environments we cover.
We need to be efficient during the information gathering/discovery stage while not taking shortcuts that could leave critical flaws undiscovered. Everyone's methodology and preferred tools will vary a bit, and we should strive to create one that works well for us while still arriving at the same end goal.
All scans we perform during a non-evasive engagement are to gather data as inputs to our manual validation and manual testing process. We should not rely solely on scanners as the human element in penetration testing is essential. We often find the most unique and severe vulnerabilities and misconfigurations only through thorough manual testing.
Sometimes the domain name can give out information about the server's purpose, for example gitlab.example.com would indicate a gitlab instance which could have public repos with leaked credentials or keys in old commits or current ones...
Enumerating one of the hosts further using an Nmap service scan (-sV) against the default top 1,000
(If we were in a medium to large-sized environment, this type of enumeration would be inefficient. It could result in us missing a web application that may prove critical to the engagement's success.)
Using EyeWitness
As mentioned before, EyeWitness can take the XML output from both Nmap and Nessus and create a report with screenshots of each web application present on the various ports using Selenium. It will also take things a step further and categorize the applications where possible, fingerprint them, and suggest default credentials based on the application. It can also be given a list of IP addresses and URLs and be told to pre-pend http:// and https:// to the front of each. It will perform DNS resolution for IPs and can be given a specific set of ports to attempt to connect to and screenshot.
or clone the repository, navigate to the Python/setup directory and run the setup.sh installer script.
Let's run the default --web option to take screenshots using the Nmap XML output from the discovery scan as input.
Using Aquatone
Aquatone, as mentioned before, is similar to EyeWitness and can take screenshots when provided a .txtfile of hosts or an Nmap .xml file with the -nmap flag. We can compile Aquatone on our own or download a precompiled binary. After downloading the binary, we just need to extract it, and we are ready to go.
Aquatone is currently under active development in a new fork, focusing on improvements and feature enhancements. Refer to the installation guide provided in the repository.
We can move it to a location in our $PATH such as /usr/local/bin to be able to call the tool from anywhere or just drop the binary in our working (say, scans) directory. It's personal preference but typically most efficient to build our attack VMs with most tools available to use without having to constantly change directories or call them from other directories.
Last updated