To make the best of LLM and RAG, the need for more data is required. Our next step will be focusing on enhancement of data protection and security through the integration of good frameworks such as guardrails and others in our system, while maintaining a promising performance in the task.
wow, this tool is pretty good to use, didn't expect that. I used Shell_GPT and aichat previously but they are not that smart in generating commands with gpt3.5-turbo. This one seems better because of the enhancement of RAG. However, the command guessing didn't work that well for me. but it's still under experimental, hope it can get officially released one day ：）