"Generating a PR summary from a diff" seems fine, but one of the primary responsibilities of a PR summary is to provide context that isn't captured in the code:
- What alternatives were considered?
- Was the change motivated by a regression? Or an outage?
- How was it tested?

Comments