(edited)
I have seen that other frameworks, after calling tools, will have the model process the content returned by tools and return the final result. However, currently, it seems that Vercel directly returns the content of tools and does not see any configurable options. Is Vercel's design like this, or is my understanding of the documentation not deep enough
including https://ai-sdk.dev/providers/adapters/langchain#example-langchain-agent-with-tools will return according to my instructions