Hi,
In my day-to-day work, I use Claude Code extensively for creating Python scripts. I’ve tried using it to help build things in NocoBase, but I find it doesn’t have sufficient understanding of the platform to produce robust results.
In v2, I noticed there’s an AI Employee that assists with creating collections. I also saw one for front-end development, though I wasn’t clear whether that’s specifically for building interfaces within NocoBase or for plugin development more broadly.
Some frameworks provide an llms.txt file to help LLMs understand the codebase. Ideally, I’d like to work towards a workflow something like this:
- AI Employee helps create product requirements
- AI Employee generates the schema/collections
- AI Employee builds the front-end/interfaces
- AI Employee creates workflows
- Humans test and refine
Is this feasible with AI Employees? What are the current best practices for using AI effectively with NocoBase? If there’s sufficient documentation available, another option might be creating a Claude Skill to give it the necessary context.
I’d really appreciate any guidance on how I can improve my development velocity with NocoBase.
Thanks, Shaheed