Mentor-reviewed labs: what reviewers actually flag

Reviewers spend most of their time on evidence, not syntax. Missing rationale for IAM scope expansions tops the list, followed by absent rollback notes when touching shared endpoints.

We also flag aspirational README files that do not match the deployed revision. A short activity log excerpt is enough to prove alignment.

When reviews bounce, we require a numbered response list mapping each comment to a change or a respectful disagreement. That habit mirrors cross-org workflow reviews in enterprise markets.