Clear visibility into agent activity in Confluence
An experiment in Confluence labels agent activity across content and audit surfaces, showing which agent acted and often who invoked it. It is coming soon to Atlassian Cloud.
- Coming soon
- Low impact
What it is
Experiment
Confluence is testing clearer labels for actions taken by agents. Version history, comments, contributors lists and supported Analytics views name the agent, and usually the person who invoked it. Where an action produces an audit log entry, that entry also names the agent.
Users see agent names, and usually the invoking person, in version history, comments and contributors lists, and agent avatars can show when an agent views, edits or comments in real time.
Audit log entries for agent actions include the agent, so admins can review that activity in Atlassian Administration.
Attribution covers Rovo, Atlassian Studio and Forge agents, plus external systems on Model Context Protocol v2, while Teamwork Graph CLI and Model Context Protocol v1 activity shows as External agent.
Written from Atlassian's release note and checked against it automatically. Atlassian's original stays the source of truth.
Before it lands
- Check audit logs in Atlassian Administration for agent activity.
- Tell teams where agent names will appear in Confluence.
- Review any external agent integrations still using Model Context Protocol v1.
What changed
- Added to Atlassian's release notes as coming soon
- First published by Atlassian
Atlas Bench began tracking Atlassian Cloud on 29 Oct 2026. Changes before that date are not in this history.