Describe the problem/error/question
Currently seems that there’s no way to pin the output of an AI Agent tool. I found this feature request but it seems forgotten.
From my point of view it seems a very straightforward feature and actually counter-intuitive that it’s not available even for the “Call n8n workflow tool”. I mean, we can pin data by calling this same workflow with the “Execute a sub-workflow” node, but we cannot pin when using the same workflow as a tool.
Sometimes a tool processing takes a lot of time and might consume a lot of resources, so I’m wondering if someone has an idea on how to tests this tools? Do you create a “mock” tool or something like that? Any other suggestion?