[new feature] Support automatic diff computation between sandbox and baseline API calls

[foxish]

New feature or functionality
What would you like Signadot to do?

This is part of the AI AutoTest umbrella - where we can expect a user to specify a stimulus (a set of requests to the system) and automatically compute differences between the responses received on sandboxes and the baseline - to detect potential regressions.

Additional considerations
Anything in particular to think about regarding this new feature?

• Shares some ideas with https://github.com/opendiffy/diffy