Releases: SaintsSec/Navi
Navi v0.6.0 - Dissidence
IT IS ADVISED THAT YOU DO A FRESH INSTALL AS OPPOSED TO RELYING ON THE BUILT IN UPDATER.
What's Changed
- LLM Chat fix by @hullabaloo-vincent in #87
- Turn Navi into a class by @hullabaloo-vincent in #88
- Type annotations by @hullabaloo-vincent in #89
- [FIX] pip install crashing and other fixes by @hullabaloo-vincent in #91
- Code cleanup and improvements by @hullabaloo-vincent in #90
Full Changelog: 0.5.2...0.6.0
Navi v0.5.2
IT IS ADVISED THAT YOU DO A FRESH INSTALL AS OPPOSED TO RELYING ON THE BUILT IN UPDATER.
What's Changed
- Alex/python venv by @AlexKollar in #83
- Navi v0.5.2 by @AlexKollar in #84
Full Changelog: 0.5.1...0.5.2
Navi v0.5.1 - Chipin' In
Navi v0.5 - "Integral"
🌟 Changes and updates:
Upgraded Navi Shell - The shell can now execute system commands, without breaking the flow of conversation. See more below!
Navi Scripts Upgrade - The new alias variable within the custom scripts allow for Navi to execute scripts right from the chat. Once again not disrupting the flow.
Navi Cryptex Integration - Thats right you heared it here. Cryptex is officially part of Navi as an example of how powerful its integration can be.
Wiki Re-write - With new power comes new documentation - Still needs more. Navi Wiki
Llama3 Integration - We made the leap from GPT to the open source Llama AI on the backend - we also customized it giving Navi her own personality and lore.
A ton of back-end fixes - This one more or less speaks for itself.
⚠️ HEADS UP!:
If you have a previous copy of Navi it is a good idea to do a clean install so the automatic updater works properly going forward.
sudo rm -rf /opt/Navi
should do the trick!
✋ If you run into issues!!
Please submit a bug report in the issues section and we will get to them ASAP!
😲 Roll Credits
I want to give a HUGE thank you to @hullabaloo-vincent for all of the Amazing work he did on bringing this version of Navi to light. I am so excited to see where we take this project in the future! - I would also like to take a moment to thank all past and current developers who worked on the Cryptex project.
Full Changelog: 0.1.4...0.5
[Pre-Release] Navi v0.1.5 - "Oh Llama"
Release Update:
Heads up this is a pre-release EXPECT BUGS.
- Ollama Integration (Beginnings of a custom model)
- GPT Has been removed.
- Added new alias to @navi from the command line
- Navi is now present on the SSG discord server.
- Apple users can direct connect to the server and access Navi on their phones.
v0.1.4 - Synthetica
Release Update:
- GPT 4.0 Integration
- Fixed the Installer for CSI Linux, Ubuntu, Arch
- Added the -q argument back into it
- Listened to @tlford82 slam his hand into the drawer. It provided motivation.
- It works on my computer (Laptop, Server, CSI VM). (So it is safe to say it works on Ubuntu and Ubuntu based systems.)
- QA Check done by @tlford82 - Might or might not have something to do with sustained injury as mentioned above.
Installation:
- Clone Navi
git clone https://github.com/SaintsSec/Navi
- Navigate into the cloned directory
- Navigate into the install directory
- run the install script:
./install.sh
The installer walks you through basic setup.
You will need to setup GPT Key Instructions below
Full Instructions here
Adding your API Key to Linux
- Open Terminal.
- Edit Bash Profile: Use the command
nano ~/.bash_profile
ornano ~/.zshrc
to open the profile file in a text editor. - Add Environment Variable: In the editor, add the line below, replacing your-api-key-here with your actual API key:
export OPENAI_API_KEY='your-api-key-here'
- Save and Exit: Press Ctrl+O to write the changes, followed by Ctrl+X to close the editor.
- Load Your Profile: Use the command
source ~/.bash_profile
orsource ~/.zshrc
to load the updated profile. - Verification: Verify the setup by typing
echo $OPENAI_API_KEY
in the terminal. It should display your API key.
Note: It's imperative to keep your API key confidential. Exposing your key can compromise the security and budget associated with your OpenAI account. Remember, to run Navi, you'll need this key, so ensure you've saved it appropriately.
Navi v0.1.3 - Reloaded
Release Update:
The main change here was the reworked implementation of GPT, while SSG is in the process of building out its Envoy AI Model.
This allows us to have a working model to run in the meantime.
Adding your GPT API KEY (Linux)
1. Open Terminal: You can find it in the Applications folder or search for it using Spotlight (Command + Space).
2. Edit Bash Profile: Use the command nano ~/.bash_profile or nano ~/.zshrc to open the profile file in a text editor.
3. Add Environment Variable: In the editor, add the line below, replacing your-api-key-here with your actual API key:
4. export OPENAI_API_KEY='your-api-key-here'
5. Save and Exit: Press Ctrl+O to write the changes, followed by Ctrl+X to close the editor.
6. Load Your Profile: Use the command source ~/.bash_profile or source ~/.zshrc to load the updated profile.
7. Verification: Verify the setup by typing echo $OPENAI_API_KEY in the terminal. It should display your API key.
Navi v0.1.2 - Rasa
Release Update:
Integration with Rasa AI
We're thrilled to introduce a variety of new capabilities through our integration with Rasa AI:
- Always On: Navi now incorporates a system service that prompts Rasa to initiate upon login, ensuring that Navi is consistently available. Detailed information about accessibility will be discussed in subsequent sections.
- Dynamic Model Loading: Rasa offers the capability to seamlessly interchange AI models developed within the framework. To facilitate this, we've developed a script to streamline the process.
- Real-time Training: Rasa's framework includes tools that allow users to modify intents and entities on-the-fly, reflecting changes immediately in the ongoing model. While we're still exploring the full potential of this feature, it offers great promise for refining AI responses.
Streamlined Installation Process
- The previous installation process was quite extensive. Much of the delay was attributed to the
jackin
script, which loaded all of Navi's installation sequences. The newjackin.py
script significantly accelerates the process. - Inclusion of Pre-trained Model: We've eliminated the necessity to train Navi after every installation. We're in the process of developing a comprehensive training set and model that encompasses a wide range of cybersecurity topics.
Enhanced Accessibility
navi
reworked alias which provides users access to Navi's comprehensive interface.naviweb
is a new alias that launches the Flask web UI for Navi.- Both aliases benefit from minimal initialization time since, as mentioned earlier, Rasa operates persistently in the background.
- As a result of number three on this list. We've also eliminated loading screens for an enhanced user experience.
Command Line Argument System Enhancement
With the aim of providing more flexible control to our users, we've implemented an advanced command line argument system for Navi:
- Simplified Interaction: By integrating command line arguments, users can now quickly and directly initiate specific functions or features of Navi without navigating through the whole Navi Shell Interface.
- Quick Help: For those unfamiliar with the newly introduced arguments, we've incorporated a -nh or navi help flag. When invoked, it will display a summary of available custom commands for Navi as well as the traditional -h flag which shows navi's command line arguments and their respective functionalities.
Revitalized Wiki
To coincide with this new version of Navi, we've undertaken a comprehensive revision of our Wiki, which includes updated procedures for training Navi via the Rasa framework. You can explore the updated Wiki here.
We deeply appreciate the time you've taken to explore our recent updates. This release signifies a major leap forward with several enhancements packed within.
Should you resonate with our project's vision and advancements, we encourage you to star the repository, follow our organization, and connect with us. Your feedback and engagement are invaluable to us.
Stay Connected
Engage with our development team on Discord
Addressing Issues
Encounter any challenges with the application? Please raise an issue and we'll attend to it promptly.
Collaborate with Us
We welcome contributions, insightful feedback, and bug reports. Explore our issues page to identify areas you can impact. Prior to contributing, please familiarize yourself with our CONTRIBUTING.md guidelines.
Navi v0.1.1 - Persistence
What's Changed
General Changes
- Training section now downloads required NLTK Modules so it works on install / jack-in.
Echo Overhaul
- The Echo-Ai framework got an overhaul including new AI training models, as well as more streamlined architecture. Feel free to head over there and check it out!
Commands:
- /update
We now have the ability to update Navi from within Navi (how meta 😎).
This allows us to keep persistence with Navi's training, configuration, GPT Logs, and .mem files.
This is going to end up saving us a lot of time when we start getting larger training datasets to work with.
Since there will be no need to keep retrain navi with each new update.
As well Navi will actually tell you on boot if there has been a version update on github.
From there, you will need to type run
and then /update
to get the new patch.
As well at the end you will be prompted to restart Navi for changes to take effect.
Take a look at the demo video below!
NaviUpdates.mp4
Future Plans:
- Ability to train Navi from within Navi
- Expanded chip market functionality.
- Cryptex integration
- Leaning more into the ML potion of Navi.
Navi v0.1.0 - First Contact
What's Changed
Commands:
- /chipmarket
Navi community scripts(chips), personality and training repo. - /pdf
New PDF export module allows for better use across the system.
Other Changes:
- GPT logging
- CSI Linux custom installation
- Typewriter function now works within more sections of Navi
- Navi chip market open to public
- Navi now offers explanations of what goes on during training in the main (non-csi) training script
- The above mentioned explanations are optional
- Argument system is live!
navi -h
Future Plans:
- Ability to train Navi from within Navi
- Expanded chip market functionality.
- Cryptex integration
- Leaning more into the ML potion of Navi.