hey all, im keen to learn from others on this issue ive been facing. I have the agent deployed in Google Agent Engine, the agent uses RAG as part of its process. Now I have a pipeline where I drop in a file and it will create a RAG corpus for me and do some basic filtering for me, once this is done the agent is called and it has to use the RAG id that was automatically created. Ive tried injecting the RAG ID into the state session but not having much luck, the Agent is using the last hardcoded RAG ID i had before deployment. how can I get the agent to grab the new RAG ID. code for how I create a new session: session = await self.remote_app.async_create_session( user_id=user_id, state=['rag_corpus_id':corpus_id']) and in my rag tool i do something like this: corpus_id_number = tool_context.state.get("rag_corpus_id")

We’re currently facing an issue with authentication popups using custom FastAPI endpoints. Our setup involves handling OAuth-based authentication through a popup window. The flow itself works correctly — the user is redirected to the provider’s auth page, signs in successfully, and the callback is supposed to hit our custom FastAPI endpoint to complete the authentication. However, the problem occurs after the authentication step: - The internal redirect URL isn’t working as expected. - The authentication status remains stuck in a pending state. - It seems like the popup isn’t properly receiving or handling the final redirect back to our app once the token is generated. We’ve already confirmed that the OAuth provider side is fine, and when we test the same redirect URL without the popup flow, it works correctly. The issue only appears when using the popup-based authentication with our custom FastAPI endpoints. Has anyone run into a similar issue or found a reliable way to handle popup-based OAuth callbacks and redirects in FastAPI? Any suggestions or best practices for managing this kind of flow would be really helpful.

Seeking assistance to scrape a website called SoFifa and no idea to handel scraping it entirely. Wish to receive help to scrape all data of it.

I have a workflow that has to do with youtube, can anyone help to debug it?

Hello, I am following https://github.com/bhancockio/adk-mcp-tutorial/tree/main/local_mcp. Upon running "python3 local_mcp/agent.py" - it did not called server.py and exited without any action. Can you help? I am on mac and using the same version for adk and mcp as mentioned in the requirements.txt