# Rolling it out to a team

A playbook for one person's setup that a team can copy, and what a team workspace adds on top: team connectors, and skills and instructions shared with every member ([Pricing and teams](/docs/pricing)).

## Today: each person, the same steps

Everything below works per person, on a free account.

1. Each person creates a 0bridge account and runs `0b setup` on their machine. [Getting started](/docs).
2. Agree on a short list of services (Linear, Slack, Notion, the error tracker). Each person connects them with their own login, so every service keeps its own permissions and nobody shares a token.
3. Put the shared rules in the repo: an `AGENTS.md` that every coding tool reads, and skills as files. 0bridge's `0b context push` then carries your own copy of them to your chat apps.
4. Use projects for the repos. `0b project link` limits each repo's tools and secrets to that repo.
5. Keep secrets in each person's vault with production behind approval, not in a shared `.env`. [Secrets](/docs/secrets).
6. Turn on conversation history if the team wants it. It is per person, so someone's history is theirs.

## Start small

- Pick one repo and two or three people for a week.
- Decide which of the optional switches the team wants: chat-app history, agent control and webhook events in chat apps (on for new accounts; anyone turns theirs off with `0b settings off …` or by asking their AI), profile in chat apps.
- Don't switch on agent control for everyone at once. Start with one machine, one repo, and plan or edit mode.
- Write down what "done" looks like for a task an agent can do alone, and what needs a person.

## With a team workspace

On the **Team** page an admin picks the team's connectors and publishes skills and instructions. Each member then runs `0b connect --team` (or uses the Connections page) to connect the ones they don't have with their own login, and gets the team's skills as `<team>--<skill>` and its instructions after their own, in chat apps over MCP and in coding agents through `0b`. Still being built with the first early-access teams: members, audit and usage in one place. [Pricing and teams](/docs/pricing).

If you want to shape it, [join early access](/app/early-access) with your team's size, the tools you use and what you want to do.

## Questions to settle first

- Who may start agents on whose machine? Today only the person whose machine it is.
- What may a chat app see? Defaults: services, not history.
- Where do shared instructions live? In the repo, not in a person's account.