AI agent does not call sub-agent tools, stops after thinking

Questions
1

Hi everyone,
I am currently building a multi-agent workflow using the sub-agent tool. However I struggle because the orchestrator agent is not calling the sub-agents. It only thinks (in between tokens) and after the thinking process (where it explicitly says it is going to call agent 1etc…) then the orchestrator just stops there and pass its thinking as output.
I am using qwen3 as orchestrator agent at the moment…
Did you encounter similar issues ?

3

Hello

I connected the sub-agents tools to the orchestrator, and inside the sub-agent I provided tool description and system prompt. Also, I set the “user message” to “AI generated” so that the orchestrator itself prompts the sub-agent.

What permission do you refer to ?

5

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.