
Confluence now shows which agent did the work and for whom
Atlassian has published how Confluence Cloud now credits work done by AI agents. Its support page, dated November 16, says that when an agent creates, views, edits or comments on content, Confluence shows the agent and the person who asked it.
This covers Rovo, Atlassian Studio agents, Forge agents, and outside tools such as Claude, Codex and ChatGPT when they work in Confluence over MCP or the Teamwork Graph CLI.
There are two kinds of credit. When an agent works for a person, Confluence shows both names, for example "Rovo with" a teammate's name. When an agent works on its own, such as in an automation, only the agent's name shows.
The names appear in contributors and version history, page analytics, comments, notifications, the live avatars on a page, and audit logs. Tools that connect over MCP v2 show their own name. Tools on MCP v1 all show as one "External agent". Atlassian notes this is not a compliance tool and points to Atlassian Guard for that.
What it means for your work
When a page changes and you do not know why, open its version history. You can now see if an agent made the edit and who started it, so you know whom to ask.
If your team lets Claude or ChatGPT edit Confluence, ask your admin which MCP version the connection uses. On v2 you see each tool by name. For a similar setup on a whiteboard tool, see Mural lets Claude, ChatGPT and Copilot work on its canvas.