Getting Started

Getting started with the ai-team-sdlc plugin

This page shows how to adopt the ai-team-sdlc GitHub Copilot plugin in any repository. The Focus Garden evidence page shows the same steps applied to a real app.

1. Install the plugin

copilot plugin marketplace add devopsabcs-engineering/ai-team-sdlc
copilot plugin install ai-team-sdlc@ai-team-sdlc
copilot plugin list

In VS Code, add devopsabcs-engineering/ai-team-sdlc to the chat.plugins.marketplaces setting and install the plugin from the Extensions or Agent Plugins view.

2. Bootstrap the repository (optional)

Run the ait-init skill. It writes .github/copilot/settings.json, a thin AGENTS.md pointer and git-ignores the .copilot-tracking/ runtime store. It never copies the plugin's agents or skills.

{
  "extraKnownMarketplaces": {
    "ai-team-sdlc": { "source": { "source": "github", "repo": "devopsabcs-engineering/ai-team-sdlc" } }
  },
  "enabledPlugins": { "ai-team-sdlc@ai-team-sdlc": true }
}

3. Write a product brief and run the lifecycle

Put the idea in a Markdown file (for example specs/idea.md) and ask the orchestrator to run it:

copilot -p "Use the ait-sdlc-orchestrate skill. Specs: ./specs/idea.md. Run the full lifecycle and stop at the human sign-off gate." --allow-all-tools

The orchestrator creates .copilot-tracking/<run-id>/ with state.json (canonical), plan.md, tasks.md, changes.md and decisions.md, dispatches the specialist agents and runs the quality gates. It always pauses for the human sign-off before deployment.

4. Sign off, deploy and keep the evidence

After a human approves, resume the run so the deploy phase ships the approved artifact. Snapshot the finished run to docs/run/<run-id>/ so reviewers can read it on GitHub.

Note

The /product-* prompt files are VS Code workspace prompts. In the Copilot CLI use the matching ait-* skills instead, for example "Use the ait-product-design skill ...".