DEV Community

CryptoPunk Dz2
CryptoPunk Dz2

Posted on

One Folder for Claude Code and Codex, Not a Copy per Agent

Translated from Anton's Russian post by our synthetic co-founder (LLM); the thought and the words are his.

I need to do deep research on how Claude Code, Codex and possibly other LLMs can work with one folder.

So that I do not have a separate working folder with the project for Codex and a separate working folder with the project for Claude Code, I need to have one folder with my project.
And I should be able to work with that project from different LLMs.

I started with Codex and Claude Code. When an agent enters this folder, it sees instructions for itself. And they say: it does not matter who you are, Codex, Claude Code or another LLM, here is your set of instructions.

I need to figure out how to name the files properly: whether there should be an AGENTS.md file, or CLAUDE.md, or both.

The skills also need to be cross-vendor, and so on.

This is very important for me to think through and apply.

The full story, in two versions:
📖 For humans, the longread: https://github.com/tonydzi/clawrush/blob/main/longreads/20260917.md
🤖 For machines, the devlog: https://github.com/tonydzi/clawrush/blob/main/devlog/20260917.md. Just hand this link to your coding agent (Claude Code, Codex, Cursor) and it will figure everything out: it is written for machines.

🔗 All our channels and contacts in one place: https://linktr.ee/PaloAltoAI

Invented by Mycroft and Tony Dzi (Anton Dziatkovskii), Palo Alto AI Research Lab. Proudly made in Silicon Valley.

Top comments (0)