CredSLayer 0.1

Extract credentials and other useful info from network captures

CredSLayer

CredSLayer goal is to look for credentials and other useful stuff in network captures. Two modes are available, pcap scanning and active processing. The latest listens for packets on a chosen interface and dynamically extracts everything it can.

Have you heard about Pcredz or net-creds ? Well this tool pushes the boundaries even further by using Wireshark dissectors, it's therefore more accurate and reliable.

This tool is really helpful if you're doing IT security or if you want to scan your own network to see if anything critical is transmitted.

Features

Right now, CredSLayer supports the following protocols:

• FTP
• SMTP / IMAP / POP3
• Telnet
• HTTP
• LDAP
• SNMP
• MySQL / PostgreSQL
• NTLMSSP
• Kerberos coming soon...

It is also able to look for email addresses and credit card numbers.

Note: To prevent false positives and performance issues, tshark (the Wireshark underlying engine) identifies protocols based on port numbers to know which dissector to use on which packets. A few protocols however have heuristic analysis to guess the protocol on different ports than the usual ones (such as HTTP). To address this issue, CredSLayer has a --map parameter, it enables you to map a (range of) port(s) to a specific protocol. This way you will be able to sniff credentials going to a specific service on a specific port you are aware of.

Install

You need tshark installed and Python >= 3.5. On some distribution tshark in shipped within the wireshark package.

Then simply install using :

$ pip install credslayer

I recommend you use a virtualenv to prevent conflicts.

Usage

$ credslayer -h
usage: credslayer [-h] [-l INTERFACE] [-s {enable,disable}] [-f IP]
                     [-m PORT:PROTOCOL] [--debug]
                     [pcapfiles [pcapfiles ...]]

Helps you find credentials and other interesting stuff in network captures

positional arguments:
  pcapfiles             pcap files you want to analyse.

optional arguments:
  -h, --help            show this help message and exit
  -l INTERFACE, --listen INTERFACE
                        start active processing on specified interface
  -s {enable,disable}, --string-inspection {enable,disable}
                        let you specify if you want to look for interesting
                        strings (email addresses, credit cards, ...) in
                        network captures. Pretty heavy on the CPU. Enabled by
                        default on pcap files, disabled on live captures.
  -f IP, --filter IP    only show packets involving the specified IP.
  -m PORT:PROTOCOL, --map PORT:PROTOCOL
                        map a port to a protocol
  --debug               put CredSLayer and pyshark in debug mode.

Get involved

Thanks to Wireshark dissectors, it's really easy to write new protocols support, you're welcome to contribute !

Contributing doesn't necessarily mean writing code. You can simply contribute by opening new issues on Github if you spot any bug or if you would like to see something added to the tool.

CredSLayer automatically dispatch packets to the appropriate file, for example if a LDAP packet is received, CredSLayer will send the LDAP layer to the analyse function inside ldap.py.

So if you want to support a new protocol, all you have to do is create a new file named after the layer name you want to parse. You can find the layer name by looking at the Protocol column in Wireshark.

There's one exception to that. If the protocol you want to extract credentials from can be embedded in other protocol, the procedure is a bit more complex (you can look at ntlmssp.py and manager.py to understand the mechanisms involved).

Otherwise, to create a simple module, there are a few things you need to know :

• The analyse() function's parameter layer is simply the layer you want to analyse, it contains everything the Wireshark dissector were able to find.
• The analyse() function's parameter session is a dictionary that enables you to have a context between the packets, it can keep stuff you found in a packet so you can access those variables to analyse the packets that follow.
• The session has 2 attributes you must know of. The first one is credentials_being_built, it's the Credentials object you must fill when you find something interesting (username, password, hash, etc). The second attribute is credentials_list, it's a list of Credentials objects and it is automatically filled with credentials_being_built when the analyse() function returns true. But in some cases (you can find one in http.py), you might want to fill that list manually.
• You must ensure the authentication was successful before logging any credentials. The credentials being built but not validated will be automatically logged as info by CredSLayer's "garbage collector".
• If at some point of your processing you realise the credentials being built aren't valid, you must call session.invalidate_credentials_and_clear_session().

Good luck ! If you need help to understand something, feel free to contact me : shellcode33{at}protonmail.ch

Credits

• CredSLayer is based on pyshark, a tshark Python wrapper

• Wireshark for their pcaps

• moloch for their pcaps

• asecuritysite for their pcaps