Skills seem to be a huge leverage point working with agents but they remain fuzzy in how they should be constructed. I like to build my skill using the agent at the end of a messy session where I successfully accomplished a goal. Then I have the agent extract the working steps into a skill that I can easily invoke again.

On this day

2024-04-13

2 years ago

I enjoyed this article by Ken about production LLM use cases with OpenAI models. When it comes to prompts, less is more