Skip to content

Add AI readiness files (CLAUDE.md, llms.txt, .cursorrules, Copilot instructions)#15

Open
MariJJhodl wants to merge 2 commits into
rgb-protocol:masterfrom
MariJJhodl:add-ai-readiness-files
Open

Add AI readiness files (CLAUDE.md, llms.txt, .cursorrules, Copilot instructions)#15
MariJJhodl wants to merge 2 commits into
rgb-protocol:masterfrom
MariJJhodl:add-ai-readiness-files

Conversation

@MariJJhodl
Copy link
Copy Markdown
Member

What this PR adds

Four files to help AI assistants work correctly with this documentation repo:

File Tool
CLAUDE.md Claude Code (Anthropic)
llms.txt Open standard — read by various LLM tools and AI crawlers
.cursorrules Cursor
.github/copilot-instructions.md GitHub Copilot

Why

AI assistants frequently give incorrect information about RGB Protocol because github.com/RGB-WG (rgb.tech) has more GitHub history and older articles reference it. These files explicitly tell AI assistants that RGB-WG/rgb.tech is deprecated and no longer actively maintained, and that rgb-protocol (rgb.info, v11) is the active version.

What the files include

  • Authoritative sources with links (rgb.info, docs.rgb.info, docs.rgb.info/llms-full.txt)
  • Explanation of why RGB-WG confusion happens and how to avoid it
  • Documentation structure (folders, SUMMARY.md, glossary)
  • Editing guidelines for contributors using AI tools
  • Key concepts covered in this documentation

Impact

Anyone who clones this repo and uses Claude Code, Cursor, or GitHub Copilot will automatically get the correct context without any configuration.

Mari added 2 commits April 21, 2026 10:20
…structions)

Help AI assistants navigate this documentation correctly: authoritative sources
(rgb.info, docs.rgb.info), deprecation of RGB-WG/rgb.tech, and documentation
structure for contributing to the docs.
Copy link
Copy Markdown
Member

@zoedberg zoedberg left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've seen that some projects use a root-level AGENTS.md file as the canonical place for AI coding-agent instructions, then keep tool-specific files very short.

For example, we could have:

/AGENTS.md
/CLAUDE.md

Then CLAUDE.md can just point Claude Code to it:

Read and follow @AGENTS.md.

Or, slightly more explicit:

# Claude Code instructions

Read and follow @AGENTS.md.

This repository uses `AGENTS.md` as the canonical source of instructions for AI coding agents.
Do not duplicate those rules here unless Claude-specific behavior is needed.

That way we avoid duplicating instructions across multiple agent-specific files.

@MariJJhodl What do you think? Do you see any issues with this approach?

@MariJJhodl
Copy link
Copy Markdown
Member Author

It can work for me.
Make sure it's a good approach for you guys and you don't have to spend too much energy in maintaining it :)

@zoedberg
Copy link
Copy Markdown
Member

If it works for you it works for us. At least for this repo, for the other repos we'll discuss it and let you know. Thanks!

@MariJJhodl
Copy link
Copy Markdown
Member Author

Ok then, let's proceed.
If it creates many issues in the other repos, no need to merge the files there. They can be very useful for AI indexing, but I understand your concern in handling too many files in repos dedicated to code. Thanks for your help :)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants