​
What makes a good prompt

Prompting well is a core skill—it unlocks faster results, better apps, and more creativity. Here’s how to improve:
  • Start with a clear goal. Be specific about what the app should do, who it’s for, and how users will interact with it.
  • Give context. Tell the AI what platform you’re building for (TBA), what tools you’re using (MiniKit, React, Tailwind), and what kind of experience you want to create.
  • Iterate in small steps. Don’t try to get everything perfect in one go. Run your prompt, review the output, and refine your request to get closer to your vision.
  • Use llms.txt files. These are AI-friendly docs provided by many blockchain tools. Including their contents (or linking to them) gives the AI better reference data.
  • Read your prompt out loud. If it sounds confusing to you, it’ll confuse the AI too.
  • Save good prompts. Treat them like building blocks. You’ll reuse them across projects.

​
What makes a prompt effective

  • State the goal and audience so the model knows what to optimize. This keeps answers focused on the right use case instead of generic solutions.
  • Use sections and lists to structure thinking and outputs, making it easier for the model to organize and for you to read.
  • Name users, roles, and permissions to anchor behavior and prevent gaps in access planning.
  • List core features as outcomes rather than vague ideas so results are actionable.
  • Define data entities and fields to guide consistent responses and align on what’s being stored or displayed.
  • Call out non-functional needs like security and performance so they aren’t forgotten in planning.
  • Provide tech preferences and constraints to narrow options and avoid irrelevant suggestions.
  • Specify deliverables and format so outputs are ready to use without heavy rework.
  • Phase the plan to keep scope lean and shippable. Example: Phase 1 = basic employee profiles and login, Phase 2 = payroll and payslips, Phase 3 = attendance and reviews.
  • Exclude out-of-scope items to prevent feature creep and keep the project realistic.
  • Invite assumptions when details are missing so progress continues without waiting on answers.

​
Additional Resources

Here are essential resources to support your Mini App development journey: