agentsclimarketplace

Version check

Skill Deadlymind/nanolama/skills/version-check

Verifies the current stable and LTS version and support window of a technology by deep research before it is installed, pinned, or upgraded — checking today's release status on official release notes, endoflife.date, PyPI, and npm rather than trusting a hardcoded number. Use when adding a dependency, choosing or bumping a version, starting a project, planning an upgrade, or asking which version or LTS to use for Django, Python, Node, PostgreSQL, Next.js, or any stack tool. Not for scanning installed dependencies for CVEs (see dependency-audit).From its SKILL.md

Install
npx -y skills add Deadlymind/nanolama --skill version-check

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • 24 days oldThe repository was created 24 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

5.5 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it

Version check (research the current LTS before you install)

When to use

Before adding, pinning, or upgrading any technology, and whenever someone asks "which version" or "which LTS". Versions and support windows move constantly; a number written down months ago is a guess, not a fact — this skill included.

Pattern

Never trust a hardcoded version — check today's date and look it up. Before you install or pin, find the current stable and LTS release and its support/EOL window from the authoritative source now, then pin the version that stays in support across your horizon. Prefer an LTS/maintained line for production; avoid a major whose EOL is near.

This step requires live web access — WebSearch/WebFetch, or a shell with network. If the session has neither, say so and do not guess: an offline answer to "which version" is a recollection, not a check.

The check — where to look, at your install time

  • Support / EOL windowshttps://endoflife.date/<product> (or /api/<product>.json): python, django, postgresql, nodejs, redis, valkey, nextjs.
  • Python packageshttps://pypi.org/pypi/<pkg>/json — read info.version, and requires_python / classifiers for the supported Django & Python range.
  • npm packageshttps://registry.npmjs.org/<pkg>/latest — the version field.
  • Django LTS roster (shifts ~every 2 years) → https://www.djangoproject.com/download/.
  • Node LTS line + codenamehttps://nodejs.org/en/about/previous-releases (even-numbered majors are LTS only after ~6 months in Current).
  • AWS managed runtimes lag upstream — check the EB platform history, the RDS Postgres release calendar, and the ElastiCache engine-versions doc; then confirm per account/region.
# Fast path: the bundled checker queries endoflife.date, PyPI and npm in one shot.
python scripts/check_versions.py                # relative to this skill; needs outbound HTTPS
#   installed as a plugin: python "$CLAUDE_PLUGIN_ROOT/skills/version-check/scripts/check_versions.py"

# Or confirm any single number yourself — never paste a remembered one:
curl -s https://pypi.org/pypi/django/json      | jq -r .info.version   # newest Django on PyPI
curl -s https://registry.npmjs.org/next/latest | jq -r .version        # newest Next.js on npm
curl -s https://endoflife.date/api/python.json | jq -r '.[0].cycle, .[0].eol'  # newest Python + its EOL
aws rds describe-db-engine-versions --engine postgres --default-only   # what RDS actually offers here

The checker prints raw numbers; you still apply judgment (LTS vs latest, compatibility, AWS-platform lag) — that is the "deep research", not just the fetch.

Example snapshot (2026-07-16 — an example, NOT an answer)

This is what a check looked like on one day, to show the shape of the output — do not reuse these numbers; run the check above for today's. Full dated table with sources: reference/stack-versions-2026-07.md.

  • Django 5.2 LTS is the current LTS (supported to 2028-04); 6.0 is the newest non-LTS release; 4.2 LTS ended 2026-04-07. The house Django pin still holds.
  • Python newest is 3.14; 3.10 reaches EOL 2026-10-31 — target ≥ 3.12.
  • DRF 3.17.x (a 3.16 pin is now behind), Channels 4.3.x (channels-redis ≥ 4.3, needs channels ≥ 4.2.2), Celery 5.6.3, Gunicorn 26.x, Uvicorn still 0.x.
  • PostgreSQL 18 newest; PG 14 is the oldest still-supported major (EOL 2026-11-12).
  • Node 24 "Krypton" is the Active LTS (26 is not LTS until 2026-10-28) — build/CI on an LTS line.
  • Frontend: Next 16.2, React 19.2, TanStack Query 5.10x, Zustand 5, Zod 4.4, Tailwind 4.3, next-intl 4.13.
  • AWS: EB newest Python branch is 3.14 (AL2023; AL2 platforms retiring); ElastiCache freezes Redis OSS at 7.1 and steers new caches to Valkey 9.1.

Adapt to your repo

Pin to your support horizon, not just "latest" — production favors an LTS/maintained line with a comfortable EOL runway; a leading-edge app may take the newest stable. Record the date you last verified beside the pin. Managed platforms (AWS EB/RDS/ElastiCache) gate what you can actually run, so check them, not only upstream.

Gotchas

  • "Latest" ≠ "LTS" ≠ "still supported" — a shiny newest release may be a short-window non-LTS, and an old pin may be past EOL. Check the support window, not just the number.
  • endoflife.date occasionally returns a wrong date — cross-check the vendor's own release page for anything load-bearing (for Django dates, prefer djangoproject.com).
  • AWS runtimes lag upstream and freeze forks (Redis OSS 7.1) — the version you can deploy may trail the one you can pip install.
  • A library can lag a new framework major (e.g. simplejwt predating Django 6.0) — confirm the compatibility matrix before pairing a new major with an older plugin.
  • Snapshots rot — this section included. Re-run the check.

See also

  • dependency-audit
  • ci-cd
  • deploy-aws

What ships with it: 2 files

9.9 KB alongside SKILL.md, 1 of them executable

scripts/

Keep looking

Skills are one crate of 326,852. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.