agentleFS
Sign inStart free
wilbeibiwilbeibi/wilbeibi-skills/skills/repo-eval/SKILL.md

repo-eval

Skill2 starsChanged 2 months agoMITView on GitHub

Evaluate a public GitHub repository's momentum and maintenance health. Use when asked whether a repository is healthy, growing, declining, abandoned, or well maintained, or when vetting dependencies and OSS projects. Do NOT use for code or architecture review.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run repo_eval.py OWNER/REPO.

    line 49
  2. Distinctive:

    Read the verdict (the quadrant) first, then sanity-check it against the raw metrics.

    line 51
  3. Distinctive:

    When interpreting, apply the REFERENCE.md rules: low momentum + steady maintenance + flat backlog = mature, not dying.

    line 53

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

repo-eval

Score a GitHub repo 0-100 on momentum (popularity trajectory) and maintenance (how well it is run), then drop it on a quadrant. Two axes decouple: a repo can be popular-but-abandoned or niche-but-impeccable. Read them together, never alone.

  • Momentum comes from the OSS Insight API (no auth): stargazers history, issue/PR-creator history, org breakdown.
  • Maintenance comes from gh (your existing auth): time-to-first-response, time-to-merge, PR acceptance rate, issue half-life, open-PR backlog age, release cadence, self-merge rate, contributor concentration, CI. These are computed month-by-month over a ≥12-month window (--months, default 18) and fitted to a slope, so trends — not just current levels — drive the verdict (e.g. response time creeping up = dying, flat = mature).

The script always prints the raw metrics beside the scores — the verdict is auditable, not a black box. See REFERENCE.md for every metric definition, the scoring thresholds, and the gotchas (squash-merge, bot noise, mature-vs-dying).

Quick start

# Full evaluation (needs network for OSS Insight + an authenticated gh)
python3 scripts/repo_eval.py vercel/next.js

# One axis at a time
python3 scripts/repo_eval.py vercel/next.js --momentum-only
python3 scripts/repo_eval.py vercel/next.js --maintenance-only

# Machine-readable (includes the per-month trend series under maintenance.monthly)
python3 scripts/repo_eval.py vercel/next.js --json

# Widen the maintenance trend window, or use weekly star granularity for young repos
python3 scripts/repo_eval.py vercel/next.js --months 24
python3 scripts/repo_eval.py vercel/next.js --period week

# Verify the metric math offline (no network)
python3 scripts/repo_eval.py --selftest

Steps

  1. Confirm gh auth status is logged in (the maintenance axis needs it). OSS Insight needs no auth but must be reachable (some sandboxes block api.ossinsight.io).
  2. Run repo_eval.py OWNER/REPO. If one axis is unreachable it degrades gracefully and the verdict says "Incomplete" — rerun with --maintenance-only / --momentum-only as needed.
  3. Read the verdict (the quadrant) first, then sanity-check it against the raw metrics. The score is a transparent weighted average; trust the metrics over the number.
  4. When interpreting, apply the REFERENCE.md rules: low momentum + steady maintenance + flat backlog = mature, not dying; green CI + recent commits that are all bot bumps = neglected, not healthy.

Notes

  • Zero third-party dependencies — stdlib + the gh CLI only.
  • Bot accounts (*[bot], dependabot, renovate, github-actions, …) are filtered before computing contributor, response-time, and merge stats.
  • Star/contributor totals are ratchets that only climb; this skill works in rates and trends, not totals, so a rolled-over project still reads as declining.
  • On hyper-active repos the issue/PR page cap can make the actual trend window shorter than --months; the report prints *_span_months and a ! note when that happens, and the affected trends return null rather than a misleadingly short slope.
  • For SaaS-OSS, high contributor concentration is expected (the company is the maintainer) — it is reported as risk, not penalized heavily. The license axis is out of scope; check it separately (OSI vs BSL/SSPL/source-available, and whether it changed).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free