I think the idea behind agents is to have them a bit less constrained. If you can force input/output with better prompts that might help. Or if you can reduce t

I think the idea behind agents is to have them a bit less constrained. If you can force input/output with better prompts that might help. Or if you can reduce the temperature of the underlying LLM generation, that could make it a bit more deterministic. Or switch to using a chain.