← ClaudeAtlas

contributor-pipeline-gardeninglisted

Maintenance of the contributor issue pipeline for JSONbored/loopover — closing issues that are already done but not marked so, and keeping the contributor-available backlog at its 50-100+ steady-state floor with well-scoped new issues. Runs every ~8h via the scheduled task (raised from daily on 2026-07-15 so the floor is maintained continuously, not caught up once a day). Invoke for "run the issue gardening", "audit open issues for stale/complete ones", "generate new contributor issues", or any recurring/scheduled run of this process. `reference.md` (next to this file) has the exhaustive label/milestone/template detail — read it before doing real work, not just this file.
JSONbored/loopover · ★ 18 · Data & Documents · score 71
Install: claude install-skill JSONbored/loopover
# Contributor pipeline gardening — loopover This repo runs on a **steady stream of well-scoped, correctly-labeled issues** for contributors — see `.claude/skills/contributing-to-loopover/`. Contributors can only open a PR that links a real, open, non-`maintainer-only` issue carrying a `gittensor:*` label. If the pipeline runs dry or fills up with stale duplicates, contributors have nothing real to do. This skill is the daily job that keeps the pipeline honest: **close what's actually done, keep the backlog stocked with what genuinely isn't.** Every run does two independent passes. Do the stale sweep FIRST — it changes what "the backlog is short" even means, so top-up sizing is only accurate after the sweep. ## Pass 1 — stale-issue sweep (do this first, every run) **The failure mode this catches:** a PR merges and actually finishes an issue's work, but the PR body says "Advances #NNNN" or just mentions "(#NNNN)" instead of "Closes #NNNN" — a deliberate non-closing reference when real work remains, but sometimes just a missed keyword when the work was actually finished. GitHub never auto-closes on a bare reference either way, so the issue sits open indefinitely looking like backlog when it isn't. This has already happened repeatedly in this repo (a whole "HELD tracker" milestone with 47/48 items silently done, several individual issues) — assume it keeps happening, don't assume today's backlog is clean. **Verify against synced upstream, not a stale local checkout.** Before