Configuration file confugion

[coffee-cup]

The railpack.json configuration file is difficult to use and set up due to its complex behavior. Railpack explicitly controls which filesystem parts and build steps are included in the final image, creating several usability issues:

Hidden Behavior

When you add a new build step to railpack.json, Railpack automatically includes it in the final image. This automatic inclusion isn't obvious and can lead to confusion.

Limited Flexibility

The current design makes it difficult to:

• Chain steps together effectively
• Run commands while only saving specific build outputs to the final image
• Use the tool without deep knowledge of Railpack's internals

Simply adding steps to the build plan doesn't execute them unless they're explicitly included in the deploy layers, but the automatic inclusion behavior obscures this requirement. Users need to understand Railpack's internal workings to use the configuration file effectively.

This issue aims to open the discussion on the config file format so that we can improve it.