Repo Sweep Automation Pack for Make.com — Stale Issue Reviewer, Proof Gate & Maintainer Digest Bots

Inspired by openclaw/clawsweeper, a maintenance bot trending on GitHub this week for how conservatively it triages backlog: it scans every open issue and PR on a rolling schedule, drafts an AI-authored close proposal only when the evidence is strong, and applies that proposal through a separate execution lane rather than closing anything on a hunch. Most teams don't run a hosted bot like that — they just watch a GitHub repo slowly fill with stale issues and unproven PRs. This pack translates the same conservative-triage pattern into three scenarios you build in Make with tools you already have.

Included templates
- Stale Issue Reviewer — on a weekly schedule, pulls open issues via Make's GitHub modules, filters out anything recently active, maintainer-authored, or pinned, then sends each remaining candidate's title, body, and comment thread to an AI module (OpenAI or Claude) with a strict prompt: propose a close reason only when the evidence supports it (duplicate-of, fixed-in-release, cannot-reproduce, no-response), and abstain otherwise. Every proposal — with its cited evidence and a confidence score — is logged to a Google Sheet or Airtable "Proposal Queue." Nothing is closed or commented on automatically; a human reviews the queue and approves each one
- PR Proof Gate — triggered on PR opened/labeled/synchronized, inspects the PR body and label set for the evidence your team requires (a passing CI link, a screenshot, a written repro), and when something's missing posts a single PR comment listing exactly what's absent, applies a "needs-proof" label, and pings the author in Slack. A maintainer can clear the gate at any time with an "override-approved" label, which fires a second branch that removes "needs-proof" and notifies the channel
- Maintainer Digest — runs after the other two on the same schedule, reads back the Proposal Queue and Proof Gate log, computes aging stats (median issue age, issues over your staleness threshold, PRs blocked more than N days), and sends one weekly Slack or email digest ranking the proposals and blocked PRs that most need a maintainer's attention

Every scenario is built the same conservative way as its inspiration: propose, log, and notify — never auto-close, never auto-merge.

Common use cases
- Solo or small-team maintainers with an issue backlog nobody has time to read end to end
- Open-source projects that want a consistent, written evidence standard for PRs from new contributors
- Teams that want a lightweight weekly repo-hygiene ritual without deploying and hosting a bot

Requirements
- A Make.com account (Core plan or higher, for scheduling and Data Store access)
- A GitHub personal access token or Make's GitHub app connection with issue/PR read and comment/label scopes
- An OpenAI or Claude API connection for drafting close proposals
- Google Sheets or Airtable for the Proposal Queue and Proof Gate log
- Slack and/or Gmail for notifications and the weekly digest

What you receive
You get a detailed, module-by-module build guide plus a scenario map for each of the three bots — trigger, filter logic, AI prompt structure, Data Store/Sheet schema, router branches, and label/notification steps all specified. Make.com blueprint exports are account-specific (they embed your own connection IDs), so rather than shipping a blueprint file that would need reconnecting anyway, this pack gives you the exact scenario map to rebuild each bot in your own account in under 30 minutes per template — no guesswork required.