Agent Merge Decides

[MrHinsh]

I really like Agent Merge and would like to get a better understanding of how it decides what feedback to act on and what feedabck to ignore. Not all comments or feedback are valuable, and some are completely unrelated to the intent of the Pull Request.

Does Agent Merge test each feedback or comment against the PR's intent?
Does Agent Merge test each feedback or comment against the Product's intent?

How can I configure what it looks at?