Agent Teams
Resolved 💬 9 comments Opened Feb 11, 2026 by elliotxin Closed Apr 6, 2026
💡 Likely answer: A maintainer (etraut-openai, contributor)
responded on this thread — see the highlighted reply below.
What variant of Codex are you using?
CLI
What feature would you like to see?
Can add Agent Teams function to Codex like claude code did?
Additional information
_No response_
9 Comments
Can you say more about what problem you'd like to see solved? What is the use case you have in mind?
He is talking about this here https://code.claude.com/docs/en/agent-teams
This may be a difficult challenge I think, as they have to implement a way for sub-agents to communicate with each other first.
Codex should not have command development algorithms; it should only provide instrumental capabilities.
There is a possibility for agents to perform tasks, which should be orchestrated by an external element.
My implementation of such a tool and examples of agent configuration settings are available in the repository https://github.com/pomazanbohdan/vida-stack.
Are there plans to make sub-agents communicate with each other at runtime?
Why is it marked as completed? I haven't seen Agent Teams in Codex, only sub-agents that are unable to communicate with one another only via the team leader, which is bad for the context window.
This is a good feature, it should not be marked as completed. Also, subagent and agent teams are two separate things.
why is it marked ask complete? lmaoo
The OP closed this issue as complete. If you'd like to suggest something similar, feel free to open your own enhancement request.