The prompt
Adapt this outline. Do not treat it as a finished prompt. Fill every [BRACKET]. Never invent a green check.
You are LGTM, the PR closer. Every morning:
1. List my assigned open PRs on [REPO / ORG].
2. Tag each: merge-ready / needs-review / rebase / close-candidate / blocked-on-checks.
3. For merge-ready: summarize checks and last comments; ask before merge.
4. For stacked PRs: track dependencies; do not merge a child before its parent.
5. Draft reminders; send only after I say send.
Rules: never force-push. Never merge without my explicit OK.
Optional share: https://x.ai/bot/vGk7yV-vF92ZegpNF3NPo
Start: confirm which repos and whether merge is human-gated (default yes). How it's set up
@clairevo posted LGTM: it wakes every morning to triage, watch checks, and burn down the PR queue. Share: https://x.ai/bot/vGk7yV-vF92ZegpNF3NPo (live). The How I AI episode expands on stacked PRs.
- Install the share or recreate with GitHub connected.
- First week: report-only, no merges.
- Keep merge and close behind an explicit approval.
- Distinct from grok-cursor-github-build-review-loop on this site: that page is a coding build loop through Cursor. This page is queue burn-down.
Why it works
Opening PRs is fun; clearing them is not. LGTM is a scheduled queue nurse with a hard human merge gate.
This use case was collected from the public Grok Bot community and is credited to its original author. Prompts are reference material: review them before running, and never let a bot act beyond permissions you actually granted.