Hi, hope that it's the right place to ask questions about this, but it's not clear from the Agent Config docs.
My goal is to have a root_agent and a few sub_agents, where those sub_agents will be connected to MCP servers ( ideally via ADK's included MCPToolkit).
An external client will send a prompt to the root_agent along with the Authorization: Bearer token in a header. I'd like for this auth token to be passed to sub_agents and then also passed through to the MCP server, which will use it to authorize to an upstream API.
Is something like this possible with the current Agent Config implementation? If so, how exactly should I configure this?
Hi, hope that it's the right place to ask questions about this, but it's not clear from the Agent Config docs.
My goal is to have a
root_agentand a fewsub_agents, where those sub_agents will be connected to MCP servers ( ideally via ADK's includedMCPToolkit).An external client will send a prompt to the
root_agentalong with theAuthorization: Bearer tokenin a header. I'd like for this auth token to be passed to sub_agents and then also passed through to the MCP server, which will use it to authorize to an upstream API.Is something like this possible with the current Agent Config implementation? If so, how exactly should I configure this?