What you'd like to see
Instead of writings prompts inline (in the yaml file), one should have the option to retrieve the prompt text dynamically by executing a helper command (like it is the case for credentials)
Why you'd like to see it
It would be possible to use shared prompts saved in prompt registries (like MLflow for example).
Furthermore it would make automatic prompt optimization (using optimize_anything for example).
What you'd like to see
Instead of writings prompts inline (in the yaml file), one should have the option to retrieve the prompt text dynamically by executing a helper command (like it is the case for credentials)
Why you'd like to see it
It would be possible to use shared prompts saved in prompt registries (like MLflow for example).
Furthermore it would make automatic prompt optimization (using optimize_anything for example).