Python Demos -- Passive OpenSource Recon
18 Feb 2018

pcap tricks + recon

A few years ago I gave a talk @ RIT SPARSA detailing some interesting Python libraries that provide a simplified gateway into the beginning statges of a pentest event. These are only a scraching of the surface. The rabbit hole goes deeper than this … much, much deeper.

The repo is found here:

These demos require:

python 2.7.3
Maxmind GeoLiteCity Database:
shodan API key:

The idea of this demo is to be able to capture some benign network traffic and learn how much insight could be gathered from it. So, to begin, capture some ethernet/ip traffic and and save it to a file somewhere. For my use case, I’m hitting the target’s website with a browser (make sure to turn off adblock and noscript addons for greater mileage).

Parsing Packets

Next, run:

    ./ --pcap [ your pcap here ]

Most of the heavy lifting here is done by the dpkt library written by Doug Song. The dpkt library is not entirely well documented, but there are enough examples out there to get you started. In this case, we’re just initializing a pcap Reader object and iterating over the resulting timestamp (ts) and data buffers. After that, dpkt has some helper classes to extract the relavent information (in this case, the ethernet frame). After that we can just test for the IP protocol and start peeling back the onion to extract IP addresses and data:

with open(path, 'r') as f:
    pcap = dpkt.pcap.Reader(f)
    for (ts, buf) in pcap:
            eth = dpkt.ethernet.Ethernet(buf)
            if eth.type == 2048:

A list of unique IP addresses can be output using a set within the nested loop (once src_ip and dst_ip are collected) like so:

    ## collect unique IP using a set
    if uniq_ip is None:
        uniq_ip = set({})


Then later output the completed set:

for x in sorted(uniq_ip):

IP to ASN Mapping

Once you have a list of unique IP addresses, save it to a file. This file can be passed on to the Team Cymru IP to ASN mapping service. Make sure to prepend the file with the word begin, and append end, each on a their own line so that all of the IP addresses are sent in a single TCP connection. Otherwise you’ll risk being null routed by Team Cymru.

Now run:

    ./ --ipfile [ your_ip_file.txt ]

Affectively we’re calling the command:

netcat 43 < list01 | sort -n > list02

This provides a little bit of insight into the services that are being utilized by the target


Geolocation is not a new practice. One of the most common (free) methods of geolocating IP addresses is to cross reference the MaxMind GeoLiteCity database. MaxMind has released their own Python library to interact directly with thier free downloads found here: MaxMind Geolite City Database

./ --ipfile [ip file here] --geodb [geolite DB here]

The output of this script is a colon separated list containing IP:LAT:LON. Save this to a file file. This geolocation is as accurate as the free Maxmind GeoLite database … so your mileage may vary. Next we’ll use Python to create a KML file to visualize this data in Google Earth.

./ --geodata [ geodata file here ] --outfile [ KML output file here ]

This makes a very simple KML file for Google Earth. Other info could be added to the cdata portion to make this file a bit more interesting, but I’ll leave that as an experiment for the reader.

HTTP Headers

In my case, I have some HTTP data in my packet capture. These HTTP requests were thoses that I initiated as well as those that were initiated asynchronously via JavaScript. These are often more interesting, and can illustrate different service usages or 3rd party partnerships. To parse out the HTTP requests and responses we’ll trigger on common HTTP TCP ports 80 and 8080.

./ --pcap [ pcap file here ]

Assuming that your traffic capture contains HTTP traffic (not HTTPS), you should see a list of URLS that were requested.


dpkt can also be used to extract DNS requests and responses. This has many wide ranging uses, we’ll scratch the surface here:

./ --pcap [ pcap file here ]

Assuming that your traffic capture contains DNS traffic, you should now see a list of DNS reqests and responses, along with the IP of the source and destination. This technique will often turn up obscure OCSP servers, thrid party ad partnershipts (potentially full of your target’s data) and can add potential unknown targets to the scope of your investigation. Use your foo to extract a list of unique hosts.

Host based whois

Using the hotsnames collected form the DNS queries in the PCAP we can get the information of all of the domain registrants. Typically large corporations mask much of thier domain whois info, but once in a while there are some easter eggs in there.

./ --hosts [ host file here ]

A list of registrant names, email addresses and name servers will fly by as you watch. Yippie! More fodder. Also interesting to note, outside the scope of this article, there are some legacy whois databases out there that contain whois information from a time before it was possible to mask.


The Shodan service is out there scanning the web everyday, and making this info available to security researchers. While there was a time that this service was free, it is no longer. I was lucky enough to get a couple keys back in those days. If you weren’t, make sure to put Black Friday on your calendar because Shodan typically has a sale on keys (~$20 USD).

You can use the same list of hosts to search Shodan to scrape lots of interesting info, including server version strings. Run:

./ --hosts [ host file here ]

This is going to make a call to using their web API. You should see a list of IPs and the servers they are running as well as a list of any additional domains that IP has been registered to. Extract a unique list of servers in the form of IP:SERVER and write to a file.


Once upon a time you would have been able to take your list of servers and scrape … but apparently that’s no longer kosher, because it doesn’t work anymore, and one try will evidently get you blacklisted. So, have a look at for now, and then skip to the next exercise.

./ --servers [ server file here ]


Tor demo that illustrates the use of Tor in python. You might want to reconfigure some of these scripts to use it. Requires:




Perhaps modifying the above example might benefit from the method shown here … dunno.


All content provided here is for informational purposes only. Opinions stated are mine and not those of my employer. I am not a doctor.