GitHub's AI dev environment, documented by the team dogfooding it
The user manual for Copilot Workspace, written by the same people who used it to build itself.

What it does This repo is the official documentation for Copilot Workspace, GitHub Next’s experiment in AI-native development. It’s a collection of Markdown guides covering setup, VS Code integration, troubleshooting, and even the project’s conceptual origins — essentially a handbook for collaborating with Copilot on repo-wide tasks using natural language and cloud compute.
The interesting bit The team built Copilot Workspace using Copilot Workspace, then wrote the manual about using Copilot Workspace using… presumably Copilot Workspace. The docs are refreshingly candid about known issues and include a responsible AI FAQ, which suggests they’re thinking about failure modes, not just hype.
Key highlights
- Covers VS Code extension setup, terminal/Codespaces workflows, and repository maintainer controls
- Includes a changelog and explicit “known issues” section (rare honesty for an experimental product)
- Documents the project’s conceptual origins and technical roadmap
- Feedback routed through Discord forum, with a dedicated safety email for harmful content reports
- Written by GitHub Next, the same internal R&D lab that shipped Copilot itself
Caveats
- No code in this repo — it’s pure documentation, so don’t expect to run anything
- Copilot Workspace itself is labeled an “experiment,” so features and availability may shift
- Some sections (like the conceptual overview) are abstract by design; practical value varies
Verdict Worth bookmarking if you’re already in the Copilot Workspace beta or tracking GitHub’s AI tooling evolution. Skip it if you’re looking for open-source code to adopt or critique — this is a manual, not a project.