Just a few major updates...
There is a lot to be covered in this update… From new team members to total system overhaul. As well as a revamped chips system!
Welcome to the team!
It is my absolute honor and privilege to say that Vincent Aliquo the VP of engineering for Haiku Inc a company I have a fair bit of history with has joined the Navi Dev team as Navi’s Systems Engineer! – Thank you so much for believing in the project! I look forward to seeing what we can make of Navi!
Have a little class...
This update as mentioned above brings a ton of new features and improvements to Navi. One of the main things we wanted to focus in on was making it so developers could easily build custom tools or even wrappers for existing tools into Navi.
We call this system the chip system (More on that later). To accomplish it we did a complete overhaul of Navi’s core. Namely Navi itself is now a class which will allow people to easily infuse Navi’s AI subsystem into their chips.
The main example being the included nmap chip. Which has been written from the ground up to showcase the power of Navi’s chat execution in conjunction with a fully integrated chip.
To use it simply type ‘Can you nmap scan port 80, 22, 443 on scanme.nmap.org’. Navi will ask if you want it to analyze the results of the scan giving a proper AI analysis of the scanned results. Making it easier to understand and giving more depth to the scan itself.
All of this without breaking the flow of navi’s chat. Take a look
Chippin' In... Chip chipping chips chipping chipping chips.
The newly revised Navi chip system is not only a guideline for building custom tools and tool wrappers for Navi. Its also a full developer focused ecosystem. Handing the power of Navi to the people. The new chip system still runs on github in the background. Instead of all chips being added to a central repo. Developers can now use their own repos thus keeping attribution. By simply tagging or labeling the repo with ‘navi-chips’ and giving at least one release. Will allow the Navi system to see the chip as available to download.
Searching for and Installing chips – from within the Navi | CLI application simply type ‘chips search’ to bring up a list of chips you can download. If you know the name of a chip already you can search for it or a term related to it. By adding the query in question to the end of the above prompt. In our case ‘chips search cryptex’ will bring up a search for its respective tool.
To install cryptex enter the following: ‘chips install navi-cryptex’ Navi will automatically go through the install process and and restart.
To see if its installed properly, type ‘–nhelp’ and the chip should show up in the list of installed chips. Alternatively you can also use the chips system built in ‘chips list’ to see a list of all installed chips on the system.
Other chip system functions We are not limited to just these functions either. As mentioned above ‘chips list’ will bring up a list of installed chips ‘chips <chip name>’ will bring up information on a given chip including links to the developers repo and an update notification. ‘chips <chip name> update’ will update a given chip. As well if you forget how to use the chips system simply typing ‘chips’ will bring up the chips system help.
Navi Server Updates:
We are always thriving to make Navi better and better. One such way is expanding Navi’s capabilities through hardware upgrades. As of the time of writing we have opted to install an Nvidia Tesla v100 AI card in server. Which will allow us much faster response times. Which brings us to a major topic… You know as if all of this has not been major.
Navi - An AI model for cyber sec professionals.
When we started the Navi project. It was a simple wrapper for Chat GPT. From there I had built a few tools around the chatbot. Using clunky menu systems and hacked together solutions. After a while we had the fortune of working with Foxlabs. They allowed us to use their homegrown AI model Echo.Thus officially moving into the realm of Navi being fully open source. As the project grew. We realized that maybe the better option was to use a more robust AI Model within Navi. We tested a few out and eventually settled on Llama. However, it was still someone else’s AI model.
Today, we are planning out the core of what will be the Navi AI Model. We will be using technologies like PyTorch to build our own model from the ground up. Thus bringing the dream of what Navi was always meant to be full circle. While we don’t have a timeline of when we can expect the model to release. We are announcing that it is in the works.
Closing thoughts
It is a really exciting time to be involved in AI and cyber. However, I need to take a moment to thank the SSG team and community at large. We would not be here if it was not for the people we surround ourselves with. You all put a lot of faith in me and my crazy ideas. Thank you so much. It means the world to me personally. You all are amazing.
So there we have it. Navi v0.6.0 has been officially released. I am sure there are parts of this I am missing in this blog post. So go check it out over on github.
If you want to get involved in Navi’s development or just come kick it with us feel free to join us on discord.
Much love and happy hacking!
