Conversation
|
My gut feeling is that this is needlessly wordy. AFAIK, your AGENTS.md is going to used up some of your context window. If we have a workspace referencing all or most of the PlayCanvas repos, all the AGENTS.md files are going to add up quickly. My approach is to get AI to write the skinniest AGENTS.md possible and then add to it whenever the LLM gets something wrong. See the developer-site one for an example: https://github.com/playcanvas/developer-site/blob/main/AGENTS.md |
Yep it is very long so im going to try condense it down |
- Condensed project overview and guidelines for clarity. - Updated section on editor loading and architecture, emphasizing backend rendering and module communication. - Simplified directory structure and code style conventions for better readability.
Summary
AGENTS.mddocumenting conventions, architecture, and best practices for AI agents and developers working on the editor frontend.Details
This is a documentation-only change — no code modifications. The file serves as a reference for both human contributors and AI coding agents to understand how the editor is structured and how to make changes correctly.