SSH-Snake: Automated SSH-Based Network Traversal

January 14, 20249 min read


SSH-Snake is a powerful tool designed to perform automatic network traversal using SSH private keys discovered on systems, with the objective of creating a comprehensive map of a network and its dependencies to identify the extent that a network can be compromised using SSH and SSH private keys starting from a particular system.

SSH-Snake can automatically reveal the relationship between systems which are connected via SSH, which would normally take a tremendous amount of time and effort to perform manually.

In other words, SSH-Snake performs the following tasks automatically and recursively:

  1. On the current system, find any SSH private keys,
  2. On the current system, find any hosts or destinations (user@host) that the private keys may be accepted,
  3. Attempt to SSH into all of the discovered destinations using all of the private keys discovered,
  4. If a destination is connected to using a discovered private key, repeats steps #1 – #4 on that system.

It’s completely self-replicating and self-propagating — and completely fileless. In many ways, SSH-Snake is actually a worm: It replicates itself and spreads itself from one system to another as far as it can.

Instead of manually jumping between systems with SSH keys like it’s a Super Mario game, let SSH-Snake do the work for you.

An in-depth look at how this script actually works, technical details, interesting discoveries, design decisions, benchmarking, and lessons learnt, check out this blog post.

SSH-Snake seamlessly emulates what a human adversary would do to discover SSH private keys and destinations that they can be used for. Written entirely in Bash, it operates with a minimal set of dependencies commonly available on major Linux systems: bashsshgetconfcoreutilsgetentawksortgreptrfind, and cat. Likewise, sudohostnameip, and xargs may also be used, but they are not required (and the script gracefully handles cases where they are not present). If a system is discovered without any of the required packages, it gracefully fails, alerting the user that the scan could not continue on that system because a required package was missing (and backtracks continues from the previous system).

SSH-Snake is completely fileless: after the user runs the script, it is passed to destinations’ bash via stdin and bash arguments (via SSH). No material evidence of the script exists on any of the systems scanned: the only evidence of the script running is in the process tree, and the substantial amount of invalid SSH attempts which will inevitably occur.

SSH-Snake takes a depth-first approach to discovery: once it connects to one system, it tries to connect further from that system before going backtracking.

The name SSH-Snake comes from the fact that the output of the script looks like a snake slithering up and down the network. However unlike the game Snake, SSH-Snake will not die when it bites its own tail (connects to a systems it has already scanned or is currently scanning): it will simply print how it connected there as normal, but return and not re-scan the destination (in order to avoid infinite recursion).


  • Recursively SSH from one system to another using local SSH private keys,
  • Fileless traversal and propogation/replication of the SSH-Snake script using only stdin and bash arguments to remote systems,
  • Automatic elevation of privileges to root using sudo if possible,
  • Discover SSH private key files from .bash_history entries,
  • Discover SSH private keys from commonly used files and folders,
  • Exfiltration SSH private keys as output of the script,
  • Configurable custom command execution on each system,
  • Plug-and-play modular system to discover private keys and systems,
  • Detect hosts from IP ranges, last logins, known hosts, SSH config files, and more,
  • Ability to detect when a system has already been scanned or is in the process of being scanned such that a network like A->B->C is able to also discover C->A but does not regress to A->B->C-A->B->C->A->B->….,
  • Ability to generate graphical visulizations of a network from the output of the script,
  • … and more.


SSH-Snake comes with some general settings that can be configured. These settings are documented in SETTINGS.md#general-settings.

SSH-Snake comes with a variety configurable/plug-and-play strategies (functions) which can be used to discover SSH private keys on a system and discover hosts and destinations to attempt to connect to. Sane defaults have been provided, however if you want to perform a scan as thoroughly as possible, then enabling more discovery techniques can help. If a scan is taking a long time, disabling some discovery techniques can help. With the exception of one strategy (find_ssh_keys), each of these strategies can be toggled off/on. These are documented in SETTINGS.md#configurable-discovery-strategies.

Understanding Output

The raw output of SSH-Snake contains a mix of infomation about discovered private keys, destinations, and error messages.

Other Tools

In addition to the ability to create visualizations of the network that SSH-Snake traverses, three other tools are provided. Namely:

  1. forward-lookup-host.py: Given a source host or destination, determine all of the systems that can be accessed either directly or indirectly (i.e. through a tertiary system).
  2. reverse-lookup-host.py: Given a destination host or destination, determine all of the systems that can either directly or indirectly access it.
  3. shortest-path-create-chain.py: Given host or destination A and B, determine the shortest path connecting the two.

Snake.sh vs Snake.nocomments.sh

Since the script is quite large, loading the script into a here-document (which it does automatically) causes bash to write to a temporary file (as it is greater than 65535-bytes).

To cut down on the size such that it remains 100% fileless, Snake.nocomments.sh has a version with all comments, unnecessary white-spaces, and blank lines removed. This cuts the file’s size down such that the temporary file is not created by bash.


  • IPv4 Only: Like all of the best programs, the script does not support IPv6. I can’t imagine there will be support for this anytime soon.
  • Port 22 Only: There is a general assumption that SSH is running on port 22.
  • GNU coreutils: The script relies heavily on GNU coreutils. I have not determined how much (if any) GNU-ism is used in the script.
  • find ... -readable ... is used in the script in multiple places. The -readable flag is not supported on all versions of find(1).

How do you vote?

0 People voted this article. 0 Upvotes - 0 Downvotes.

What do you think?

Show comments / Leave a comment

Leave a reply