Subagent delegation not working

Support Help
, ,
1

Describe the Bug

So I’m not sure if this is a bug or I misunderstood the docs on how subagent delegation.

I created a complex setup with multiple subagents, i can call them with /agentname but the main agent never delegates to them. When I ask it explicitly, it says it cannot and “acts” as the agent (doesn’t solve the reason I set them up in the first place, context window savings)

I then took a different repo, set up a single subagent using the example propmpt on the subagents documentation page, and ask the main agent explicitly to invoke that agent. It said it had no such ability “We can’t spawn Cursor’s subagent UI from here. Acting as the Verifier: …”

Am I doing something wrong, or is it a bug?

My ultimate goal is for the main agent to automatically spawn/delegate to researcher, coder and tester subagents

Steps to Reproduce

Create a subagent .cursor/agents/verifier.md and ask the agent to use it

Expected Behavior

Main agent delegates to subagent, as described in the docs

Operating System

MacOS

Version Information

Version: 3.2.16
VSCode Version: 1.105.1
Commit: 3e548838cf824b70851dd3ef27d0c6aae371b3f0
Date: 2026-04-28T21:07:47.682Z
Layout: editor
Build Type: Stable
Release Track: Default
Electron: 39.8.1
Chromium: 142.0.7444.265
Node.js: 22.22.1
V8: 14.2.231.22-electron.0
OS: Darwin arm64 25.3.0

For AI issues: which model did you use?

composer-2

Does this stop you from using Cursor

No - Cursor works, but with this issue

4

Hi @Reuven_Kaplan!

Cursor should be able to spin up subagents when it thinks it’s appropriate.

Can you share .cursor/agents/verifier.md here, and also what prompting you’re using that you hope will make the subagent spin up by itself? Would like to try and reproduce on my end!

Especially what you asked that caused it to say:

Normally, it should happen quite easily:

Screenshot 2026-05-04 at 14.46.50
Screenshot 2026-05-04 at 14.46.501530×370 27.3 KB

6

verifier.md.txt (2.1 KB)

This was a clean setup, first prompt was

“Create a subagent file at .cursor/agents/verifier.md with YAML frontmatter (name, description) followed by the prompt. The verifier subagent should validate completed work, check that implementations are functional, run tests, and report what passed vs what’s incomplete.”

The follow up prompt was

“can you call the verifiyer and ask it to run the tests”

8

What if you wrote:

“Can you call the verifier subagent and ask it to run the tests?”

Spelling will matter here, and helping the Agent understand it’s looking for a subagent to call can make a big difference!

9

Still no luck

"Cursor doesn’t expose a API for spawning another agent from chat; we’ll follow the Verifier playbook: discover the test command, run it, and produce the report.

I can’t invoke another Cursor subagent from here; only you can start Verifier from the UI (Composer / agents). I ran the same checks the Verifier agent describes so you still get a verification-style result."