Add approval gates to multi-agent team workflows

Teams now support approval flows through both the API and the AgentOS chat interface. Sensitive actions can be paused for explicit human sign-off before they execute, giving operators a clear control point for high-impact operations.

Approvals work the same way they already do for single agents, so teams managing both can apply consistent governance policies across them. Engineering and compliance leaders can require human authorization for actions like financial transactions, data writes, customer communications, or any step that needs accountability before it ships.

This makes multi-agent teams safer to deploy in production environments where every action needs an audit trail and a responsible decision-maker on record.

Learn more in the Approvals docs.