Currently, the entities used to implement durable agents use different method names for invoking the agents, RunAgentAsync on .NET and run_agent on Python. This inconsistency makes it difficult for tools (such as the DTS Dashboard) to interact with the agents at the entity-level (as opposed to having developers go through their application level for testing).
We should use a common set of entity method names across the platforms. (I'd suggest RunAgent but I don't have a strong opinion.)