Automatically parse feedback value


When feedback is submitted, the `actionValue.feedback` comes in as a JSON.

Should we parse this thing to help out devs? Or leave it up to them? I see a case for both.
If we parse it for them, improved devex
If we don't, they can store the whole thing, and if more fields get added, they won't need to change anything.
Though, I think most people are parsing it and storing the text directly...

Thoughts?

Start by tracing how submitted feedback reaches actionValue.feedback and how developers currently consume that value. Compare the implications of parsing it with preserving the JSON, using the issue's stated compatibility concerns as acceptance criteria. Done requires a decided behavior and validation of the resulting feedback shape.

Written by the indexing model from the issue text.