Native WinPC Development such as Game Development & Codex least friction

I’m working on Windows PC applications - such as game development - using Codex as an extension to VS Code (or Cursor, etc.) and the docs for the Codex extension state:

Windows support is experimental. For the best Windows experience, use Codex in a WSL workspace

Though then we’re also essentially developing a Linux app and Windows PC tooling can have issues with the files on the virtual Linux drive. All the friction moves to the Windows development tooling.

I am thinking to try to use the VS extension without WSL and just adding a bunch of Linux style command line apps to mimic Linux like behavior so that at least reduces some Codex friction. Maybe even combine this with an agent skill windows-cli-map such as Use rg instead of grep, etc.

Anyone have any best/better practices for using Codex to target Windows PC builds?

Here’s hoping for Codex/OpenAI native Windows tooling in the near future. :crossed_fingers:

4 Likes

I have since tried this - got Codex to recommend a bunch of Windows CLI equivalents / replacements for 'nix style CLI apps and made their recommended use into a skill. Seems to work great!

3 Likes

Hey @Noolarch

Welcome to the community.

It seems like a lot of windows users could benefit from this project. Do you plan to open source the tools/skills in future?

1 Like

Yes - making a Windows CLI helper skill open source would be a good idea - esp. if others evolve it even more.

Is there a recommended Skill repository to use?
I was doing this for Codex though it should work for any model.

3 Likes

I can not recommend a specific repository.

I would suggest using a GitHub account.
At present OpenAI GitHub is showing unicorns, so can’t link to an example.

1 Like

Hi,

Apologies for the delayed response.

Following up on your question: at the moment, the best option is to publish the skill in a GitHub repository. This allows others to install it directly using the GitHub URL, for example:

$ skill-installer install https://github.com/openai/skills/tree/main/skills/.experimental/create-plan
2 Likes