regit / pshitt

Passwords of SSH Intruders Transferred to Text

PSHITT

Introduction

pshitt (for Passwords of SSH Intruders Transferred to Text) is a lightweight fake SSH server designed to collect authentication data sent by intruders. It basically collects username and password used by SSH bruteforce software and writes the extracted data to a file in JSON format.

pshitt is written in Python and uses paramiko to implement the SSH layer.

Installing pshitt

From Python Packaging Index (PyPI) using pip :

pip install pshitt

Install from source :

git clone https://github.com/regit/pshitt.git

NOTE: if you are installing from source, make sure you install paramiko and python-daemon packages.

Running pshitt

If you installed via pip :

./pshitt -o passwords.json

If you installed from source, go into the source directory and run :

./pshitt.py -o passwords.json

This will run a fake SSH server listening on port 2200 to catch authentication data sent by the intruders. Information about SSH connection attempt will be stored in the passwords.json using JSON as format :

{"username": "root", "src_ip": "116.10.191.184", "password": "P@ssword", \
 "src_port": 41397, "timestamp": "2014-06-25T21:35:21.660303"}

Full options are available via '-h' option :

usage: pshitt [-h] [-o OUTPUT] [-k KEY] [-l LOG] [-p PORT] [-t THREADS] [-v]
              [-D]

Passwords of SSH Intruders Transferred to Text

optional arguments:
  -h, --help            show this help message and exit
  -o OUTPUT, --output OUTPUT
                        File to export collected data
  -k KEY, --key KEY     Host RSA key
  -l LOG, --log LOG     File to log info and debug
  -p PORT, --port PORT  TCP port to listen to
  -t THREADS, --threads THREADS
                        Maximum number of client threads
  -v, --verbose         Show verbose output, use multiple times increase
                        verbosity
  -D, --daemon          Run as unix daemon

Using pshitt data

As the format is JSON, it is easy to use the data in data analysis software such as Splunk or Logstash.

Here's a sample configuration for logstash :

input {
   file {
      path => [ "/var/log/pshitt.log" ]
      codec =>   json
      type => "json-log"
   }
}

filter {
    # warn logstash that timestamp is the one to use
    if [type] == "json-log" {
        date {
            match => [ "timestamp", "ISO8601" ]
        }
    }

    # optional but geoip is interesting
    if [src_ip]  {
        geoip {
            source => "src_ip"
            target => "geoip"
            add_field => [ "[geoip][coordinates]", "%{[geoip][longitude]}" ]
            add_field => [ "[geoip][coordinates]", "%{[geoip][latitude]}"  ]
        }
        mutate {
            convert => [ "[geoip][coordinates]", "float" ]
        }
    }
}

output {
  elasticsearch {
       host => "localhost"
  }
}

Basically, it is just enough to mention that the pshitt.log file is using JSON format.