Multi-agent programming

Ideas Feature Requests
,
1

Feature request for product/service

Cursor IDE

Describe the request

Hi everyone, I’ve been watching the multi‑task mode for a while now, but it’s not quite what I expected.
Maybe I don’t understand something or I don’t know Cursor well, but in my view, it would be great if it were possible to create a “Team of AI agents”. Assign specific roles and instructions to each of them, and choose the right model for each of them.
And most importantly, these agents shouldn’t be “one‑day wonders”. They should be able to maintain long‑term interaction.
For example, in web development, you can create a team of agents that includes a team lead (architect), a frontend developer, a backend developer, a security specialist, and a reviewer. And these agents shouldn’t be one‑off roles. They should be able to communicate with each other, consult, and so on.
Yes, there’s the issue of context length, but nevertheless, all of this can be solved with the right system instructions.
The main agent — the conductor — is a powerful model with a good context; others can be assigned to cheaper models.
To be honest, I’ve been dreaming for a long time about having a full‑fledged team of AI agents who work long‑term, can interact with me, and I can give instructions both to their director and personally to each agent in multiple iterations in the chat.
It would be great to see something like that in the long term.
Maybe something similar already exists, but I don’t know about it. Maybe this topic has already been raised here, but in any case, I’d be glad to hear your feedback!

Operating System (if it applies)

Windows 10/11

5

Feature request for product/service

Cursor IDE

Describe the request

Wish: a single shared chat where a human and multiple agents (different models) discuss together — same thread, same context, visible to each other. Sequential “ask model A, paste to model B” loses the point. Use case: architecture review with two reviewers plus the implementer in one conversation.

8

Hey Stas, thanks for the detailed feature request.

Some of what you described already works. Cursor has custom subagents. You create files in .cursor/agents/*.md, give each one its own role and instructions, and pick a model for each one, like a cheaper model for a reviewer and a stronger model for an architect. Then the main agent can delegate tasks to them. Your team setup idea frontend, backend, security, reviewer can be done this way. Guide: Subagents | Cursor Docs

What’s not available out of the box yet is a truly long-lived team of agents that keeps working together across many iterations under one conductor and can message each other. The team is actively exploring this, so keep an eye on release updates.

Related threads where you can add your vote: