Skip to main content

PyPI-first scaffold for a dual Codex and Claude science workspace.

Project description

scix

scix creates a science-oriented workspace that is ready for both Codex CLI and Claude Code. You install scix once, run scix up inside a new empty folder, and it builds the workspace for you.

What scix does

scix sets up:

  • a shared AI policy system for Codex and Claude
  • a repos/ folder for cloned reference repositories
  • a workspace/ folder for your own experiments and notes
  • a local Python environment named xenv/
  • generated AGENTS.md, CLAUDE.md, skills, hooks, and tool configs

Who this is for

This README assumes:

  • you are new to AI tools
  • you are new to the Terminal
  • you are new to VS Code

Before you begin

You need:

  • a macOS or Linux machine
  • a GitHub account
  • a PyPI-accessible Python install that can run pip install scix
  • Codex CLI installed if you want to use Codex
  • Claude Code installed if you want to use Claude

If you are using macOS and have never opened the Terminal:

  1. Press Command + Space.
  2. Type Terminal.
  3. Press Return.

Quick start

Create a brand new empty folder, then move into it:

mkdir my-scix-work
cd my-scix-work

Install scix:

pip install scix

Run the setup command:

scix up

scix will ask you to confirm that:

  • the current directory should become your scix workspace
  • this directory is where you want to do your scix work

If the folder is not empty, scix up stops by default. That is intentional.

What scix up will do

It will:

  1. create the scix workspace files
  2. install pyenv if needed
  3. install Python 3.11
  4. create a local environment named xenv/
  5. install scix and the five science packages into xenv/
  6. clone the five reference repositories into repos/
  7. generate Codex and Claude config files

Activate the local Python environment

After setup finishes, activate the environment with:

source xenv/bin/activate

When it is active, your Terminal prompt usually changes. From then on, Python and pip commands use the local environment inside this workspace.

Log into Codex

Open a Terminal in your scix workspace and run:

codex login

If you are using an API key instead of the normal login flow:

printenv OPENAI_API_KEY | codex login --with-api-key

To check whether you are already logged in:

codex login status

Log into Claude

Open a Terminal in your scix workspace and run:

claude auth login

If you use a long-lived token flow:

claude setup-token

To check whether you are already logged in:

claude auth status

Use VS Code

If Visual Studio Code is installed, open it and choose:

  1. File
  2. Open Folder...
  3. Select your scix workspace folder

Then open the built-in terminal in VS Code:

  1. Terminal
  2. New Terminal

Activate xenv there too:

source xenv/bin/activate

If the code command does not work in the Terminal, open VS Code and run:

  1. Command Palette
  2. Type Shell Command: Install 'code' command in PATH
  3. Press Return

After that, you can open the current folder from Terminal with:

code .

Start Codex or Claude inside VS Code

From the VS Code terminal:

codex

or:

claude

Important folders

  • repos/: cloned reference repos such as kintera and pydisort
  • workspace/: your own experiments, notes, notebooks, and rough work
  • xenv/: the local Python environment for this workspace
  • ai/: the shared rules and templates that generate Codex and Claude files

Helpful commands

Regenerate all Codex and Claude files:

scix sync

Check whether the workspace is healthy:

scix doctor

Clone any missing reference repositories again:

scix install-repos

Notes for source-repo developers

This GitHub repository is the source of truth for the scix package. End users normally do not clone it. They install from PyPI and run scix up.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

scix-0.1.0.tar.gz (18.8 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

scix-0.1.0-py3-none-any.whl (23.3 kB view details)

Uploaded Python 3

File details

Details for the file scix-0.1.0.tar.gz.

File metadata

  • Download URL: scix-0.1.0.tar.gz
  • Upload date:
  • Size: 18.8 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for scix-0.1.0.tar.gz
Algorithm Hash digest
SHA256 f75f8bc1f734d392fc2f347cffc7d5faebddf6f2437339ddc3933416c82fd96c
MD5 80dcd49b3260e399669fcc3478b5e916
BLAKE2b-256 6644d652b9b2cbdfb645ec4c089262028ff39b728c81bbce91236cc74bd8f758

See more details on using hashes here.

File details

Details for the file scix-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: scix-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 23.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for scix-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 47c03a0bc79e982cae597c2c6c7cb5219ee8ef5ead342d6a09b4a1e342c9d387
MD5 fe87da9252154536884b34fac3e5245d
BLAKE2b-256 3bafdad940d0e6d8a17b08dca37ef5a1125981284be665c5989ddb170a8c432a

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page