Hey, great request, and it’s awesome that you want to dig deeper instead of just getting ready-made code. Here’s where I’d start.
On Cursor itself and its core concepts (agents, rules, context, tools):
- Quickstart: Quickstart | Cursor Docs
- How the Agent works: Overview | Cursor Docs
- Help Center on agentic coding, it explains the mental model really well: What is agentic coding? | Cursor Docs
- Context (how the model “sees” your project) and rules: @ mentions and context | Cursor Docs and Rules | Cursor Docs
For your use case (research plus data processing automation), it’s also useful to read other users’ experience:
- Cursor for Data Science: Cursor for Data Science - my experiences
A practical way to learn: use Cursor itself as a tutor. Open Ask mode (the mode where the model answers and explains, but doesn’t edit files) and ask it to break down every step, like “explain what this code does and why,” or “what is an agent and how is it different from normal autocomplete.” This usually makes the concepts click faster than dry theory.
If you describe what kinds of data and tasks you work with, I can give more targeted advice.