Refactor commenting strategies for better reusability

[verbanicm]

TL;DR

Current commenting strategies in the plan and apply commands duplicate code frequently.

Detailed design

Two options top of mind:

1) We can create helper functions to generate the different types of messages (e.g. https://github.com/abcxyz/guardian/pull/352#discussion_r1704680358)

2) We can create a reporter/commenter interface and implement it for github to allow for potential future strategies. The interface would require methods for generating the different messages and writing them to the output (github, stdout, etc)

Alternatives considered

No response

Additional information

No response