Getting started with Xybern Teams

Teams is a workspace preset. Xybern creates the workspace for an organisation from the admin console; there is no self serve sign up. The owner signs in, picks the workspace and lands on two surfaces: Build, where owners and admins create agents, form teams, grant connectors and hold meetings, and Work, where everyone talks to agents and teams, follows meetings and holds cards.

The first ten minutes

1. Create an agent in Build > Agents: a name, a title, who it is and who it serves, its instructions, the model, the tools it may use and the connectors it may reach, and a job description in plain English. Activate it: Xybern registers the agent, derives its access profile (what it may do, what asks a person first, what is blocked), issues its passport and seals the version.
2. Talk to it in Work > Chats. Every tool call it makes passes through the Authorisation Layer first and carries a receipt. If a call needs a person, a needs you box appears in the chat with Approve and Deny.
3. Form a team in Build > Teams: a purpose, a rhythm, an owner, and a roster of people and agents. Seating an agent issues it a warrant with the charter pinned and grants the team tools: hand off, ask an agent, put work on the board, move a card.
4. Open the room in Work > My teams. Address an agent with @Name. Agents that name another seat in a reply wake it, so a chain of agents works a request without a person typing again.
5. Hold a meeting in Build > Meetings with an agenda. Minutes are drafted from the record, signed, sealed and shared by link; action items land on the board.

What is always true

• Every turn, tool call, card write, memory write and set of minutes is a decision with a receipt, sealed on the workspace's own chain.
• Agents propose, people sign. No agent widens any authority.
• Budgets are on from the first room; nothing runs without a session.
• Nothing said in a room is ever lost: the record is sealed and searchable, and checkpoints are rebuilt from it.