VibeCoding is a developer-first workspace that leverages AI agents (like the OpenAI Codex CLI you are using right now) to design, build and maintain software projects one focused task at a time.
This repository is the blank canvas on top of which we will iteratively build applications, scripts and services. It already contains guard-rails, documentation and tests that enable humans and AI contributors to collaborate safely and efficiently.
Instead of shipping a monolithic boiler-plate, VibeCoding starts with only the scaffolding that ensures quality and order:
- A clear, up-to-date ROADMAP that explains where we are heading.
- A living TODO list that breaks the roadmap into small, bite-sized tasks.
- Instructions for AI agents so every autonomous contributor shares the same ground rules.
- HUMAN.md records manual steps or environment requirements that require a human touch.
- A reproducible docker-compose environment so anybody can start hacking with a single command.
- Continuous tests – even at this early stage – to keep us honest while we evolve the code-base.
- Source code is organized by language under
src/<language>/.
Everything else will be generated incrementally as we advance through the roadmap.
# Clone and enter the repo
git clone <your-fork-url>
cd vibeCoding
# Spin up the dev stack (adds a dedicated network, volumes, etc.)
docker compose up -d
# Run the test-suite
docker compose exec app pytest -qPull-requests are welcome! Please read docs/CONTRIBUTING.md for guidelines.
If you are an AI agent (👋), read INSTRUCTIONS.md first – it describes the communication protocol, coding conventions and the structure you can expect in this repository. Consult HUMAN.md for any tasks that require human intervention.
The project is in bootstrap stage. The immediate goals are listed in TODO.md and the higher-level plan sits in docs/ROADMAP.md.
This project is licensed under the MIT License.
Made with care by the VibeCoding community.