loop-review
Loop Review
Two loops, in this order, on each pull request: review until a pass is empty, then simplify until a pass is empty. Then Autopilot: run /autopilot until the pull request is merge-ready.
One invocation owns the whole surface. Finish every sibling of every condition in this run. The next invocation is not where a name on this run's list is allowed to appear. A type or spelling the statement does not contain is not on that list. The exception is a list whose job is to check or supply every member of a set (Sibling walk): every member of that set is on the list, named in the statement or not.
Usage
/loop-review <PR> [<PR> ...]
Each <PR> is a GitHub pull request URL, a GitLab merge request URL, owner/repo#number (GitHub) or group/project!number (GitLab), separated by spaces or newlines. If none is given, stop and show the usage line. Do not guess a pull request. Drop duplicates, and list anything that is not a pull request as skipped in the batch report. scripts/setup.py prints each one's URL as PR=; every later <PR> is that URL.
<skill-dir> below is the directory this SKILL.md was read from.
Batch
With one pull request, run everything in this conversation. Do not start a subagent: it would read this skill and both review skills again into a cold context.