Use this optional skill for work that is already implemented, not for planning or executing new work.
If trust, the exact change to review, or the comparison range is unclear, stop and ask before fetching GitHub, GitLab, or Git data.
Read review-core-guidance.md first.
Then choose the review mode and evidence source together, and load only the needed guidance files. Load both mode files only when the user wants both outputs:
- Assessment — pick this when the user wants a verdict, recommendation, PR or MR review file, or reconstruction of the change after the fact. Load assessment-guidance.md.
- Walk-through — pick this when the user wants file ownership, review blocks, reading order, or inline file-wise explanation. Load walk-through-guidance.md.
If the review target is a specific GitHub or GitLab pull request or
merge request page, number, or other remote review target, also load
github-gitlab-evidence-guidance.md.
Do not load github-gitlab-evidence-guidance.md for branch diffs,
commit ranges, current local changes, or other local Git review
targets.
Do not load spec-loop-plan-task by default. Only the assessment mode
reuses parts of it, and only for the specific conventions it actually
needs.
Use spec-loop-clarify-task when an unresolved choice could materially change trust, evidence source, comparison range, review mode, document layout, or verdict expectations.