Optimizing GitHub Copilot for Rhino: Special Prompts for R Shiny Developers

R Shiny developers face a unique challenge in 2025: while AI coding assistants have revolutionized development workflows across most ecosystems, R Shiny applications, particularly those built with Rhino, remain underrepresented in the training data of large language models. While this has been substantially improved with the recent update to the Shiny extension for VS Code presented on the ShinyConf 2025, support for Rhino is still lacking.

At Appsilon, we recognize this gap and have addressed it directly in our latest Rhino release.

New to Rhino? Start here with our Rhino Introduction Guide

The Problem: LLMs Don't "Speak" Rhino Fluently

If you've tried using GitHub Copilot or Cursor with Rhino projects, you've likely noticed the AI struggling to generate appropriate suggestions. Maybe it suggested using library instead of box::use, or it didn't understand how to structure a module correctly. This is because these AI tools have been trained on a wide variety of codebases, but Rhino's unique patterns and practices are not as common in the datasets they learned from.

The result? Less accurate code completions, misunderstood architecture patterns, and ultimately, a diminished developer experience.

Our Solution: Custom Prompts for Better AI Assistance

Rather than waiting years for these models to naturally encounter more Rhino code, we've developed specialized prompts that effectively "teach" these AI tools how to better understand and generate Rhino-compatible code.

We've released a comprehensive set of tailored instructions for GitHub Copilot and Cursor that help these tools understand:

• Rhino's directory structure and file organization
• Proper module implementation patterns
• Best practices for box functions
• Testing conventions

How to Use These Prompts

We've documented everything you need in our updated Rhino documentation.

For developers who want to contribute or customize these prompts further, we've created a dedicated repository at Appsilon/rhino-llm-rules. We encourage you to leave us feedback and suggest improvements.

The Impact: More Productive R Shiny Development

By bridging this knowledge gap, we're enabling Rhino developers to enjoy the same productivity benefits from AI coding assistants that other frameworks have experienced. The specialized prompts help you:

• Generate boilerplate code faster
• Receive contextually appropriate suggestions
• Maintain Rhino's architectural patterns consistently
• Reduce the learning curve for new Rhino developers

Before and after comparison. Note that the version without the custom instructions uses library instead of box:use and doesn’t suggest a place where to store function file, logic directory.

We Need Your Feedback

These prompts are just the first step. As you use them in your Rhino projects, please share your experiences and suggestions in the repository issues section. Your feedback will help us refine these instructions to make AI tools even more helpful for R Shiny development.

Visit our GitHub repository to contribute your insights or report issues with the current prompt set.

Looking Forward

As we continue to evolve Rhino for production-ready Shiny applications, we're committed to ensuring it works well with modern development workflows, including AI-assisted coding. These specialized prompts represent our pragmatic approach to improving developer experience without waiting for the broader AI ecosystem to catch up.

Try the prompts in your current Rhino project and let us know how they work for you.

See Rhino in Action. Watch our ShinyConf 2025 presentation on the latest Rhino updates