agentleFS
osmlabosmlab/name-suggestion-index/.claude/skills/checkup/SKILL.md

checkup

Skill893 starsChanged yesterdayBSD-3-ClauseView on GitHub

Check up on a repo after time away — where things stand, whether it still builds, security alerts, outdated dependencies, sync manifest drift, and what a newer model would do differently. Reports a numbered list; makes no edits. Use when returning to a project after a while, or when asked for a checkup or health check.

8 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use report.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • MicrosoftSkill

    report

    Turn an exploration (threads, findings, charts) into a single Markdown report — note, blog post, executive summary, KPI dashboard, slide brief, or multi-section analytical report, with embedded charts.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

9 rules. 0 are common, 3 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

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

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

    If the user gave a focus with the prompt, use it.

    line 11
  2. Distinctive:

    Before running anything, recover context:

    line 17
  3. Very specific:

    Read AGENTS.md, the project's knowledge (knowledge/current.md and knowledge/backlog.md, see AGENTS.md), a root PRIVATE.md if there is one, …

    line 19
  4. Very specific:

    If gh is available: open PRs and issues (gh pr list, gh issue list), and recent CI runs on the default branch (gh run list --branch …

    line 23
  5. Very specific:

    Run the project's check, build, and test scripts from package.json (often bun run all).

    line 32
  6. Distinctive:

    Run git status again afterwards.

    line 33
  7. Very specific:

    If gh is available, GitHub's open alerts: gh api 'repos/{owner}/{repo}/dependabot/alerts?state=open', and the same for code-scanning/alerts …

    line 40
  8. Very specific:

    Read the local .sync.manifest.md at the repo root.

    line 58
  9. Distinctive:

    Start with a short Where things stand paragraph: last-activity date, branch and working-tree state, CI on the default branch, and whether …

    line 88

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “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

Give this repo a checkup after time away: find out where things stand, what has drifted while nobody was looking, and what you'd do differently now. This is about drift since the last round of work, not a general code review (that's /suggest).

Report only. Do not edit files, upgrade dependencies, sync, or commit. Installing from the existing lockfile and running the existing build and tests is fine. The user picks which items to act on.

If the user gave a focus with the prompt, use it. If a check can't run (no network, no gh, a missing tool), say so in the report and move on.

This skill assumes a Bun project. For other ecosystems, use the equivalent commands (see Other ecosystems).

1. Where things stand

Before running anything, recover context:

  • Read AGENTS.md, the project's knowledge (knowledge/current.md and knowledge/backlog.md, see AGENTS.md), a root PRIVATE.md if there is one, and skim the README.
  • git log -1 --format='%ad %s' --date=short gives the last-activity date. It anchors section 6.
  • git status, git stash list, unpushed commits (git log @{u}.. if there's an upstream), and local branches not merged into the default branch.
  • Leftover branches on the remote. List them with git ls-remote --heads origin, not git branch -r: remote-tracking refs keep branches that were already deleted on GitHub until a git fetch --prune.
  • If gh is available: open PRs and issues (gh pr list, gh issue list), and recent CI runs on the default branch (gh run list --branch <default> --limit 5).
  • Open items in current.md and backlog.md. Some may be easy now, or obsolete.
  • knowledge/ files over the size budgets in knowledge/README.md. Suggest /reflect to prune them.

2. Baseline health

Find out whether the project works as it is, before anyone upgrades anything:

  • bun install --frozen-lockfile. A failure here (lockfile out of step with package.json) is itself a finding.
  • Run the project's check, build, and test scripts from package.json (often bun run all). Skip anything that deploys, publishes, or starts a long-running server.
  • Run git status again afterwards. If the build changed tracked files, report it.

Record pass or fail, and for failures, the first real error. If the baseline is red, read everything else in that light.

3. Security

  • bun audit for known vulnerabilities in installed dependencies.
  • If gh is available, GitHub's open alerts: gh api 'repos/{owner}/{repo}/dependabot/alerts?state=open', and the same for code-scanning/alerts and secret-scanning/alerts. A 403 or 404 usually means the feature is off or the token lacks access; note it and move on.
  • For each vulnerability: package, severity, direct or transitive, whether it reaches shipped code or only dev tooling, and the fixed version.
  • For secret-scanning alerts, give the alert number, secret type, and location. Never repeat the secret value in the report.

4. Dependencies and tooling versions

  • bun outdated. Group patch and minor updates (usually safe together) apart from major ones (each needs a look at its changelog), and name the major bumps that matter most.
  • Deprecated or unmaintained packages, and packages Bun or the platform now handles natively.
  • Runtime: the installed bun --version, engines.bun in package.json, and the latest Bun release.
  • GitHub Actions: compare each uses: pin in .github/workflows/ with the action's latest release (gh api repos/{owner}/{action}/releases/latest --jq .tag_name).

Check live sources for every version number (the registry, gh, release pages). Don't trust memory: even a newer model's knowledge stops before today.

5. Sync status

Report what /sync would change, without doing it. (If this repo is agent-practices, skip this section.)

  1. Fetch the source manifest: https://raw.githubusercontent.com/rapideditor/agent-practices/main/templates/.sync.manifest.md.
  2. Read the local .sync.manifest.md at the repo root. If it doesn't exist, say the project hasn't been synced yet and recommend /sync.
  3. For each file in the source manifest, compare versions (missing row = version 0). Use the per-file instructions to tell files that are behind apart from files that don't apply here (e.g. tsconfig.json in a project without TypeScript).
  4. Also note orphaned rows (in the local manifest, gone from the source) and legacy .github/prompts/*.prompt.md files still present.

Make this one item that lists what's behind (local → canonical version) and recommends /sync, not one item per file. Show each file's local Notes next to it: they record deliberate differences from the template, so that drift is expected, not a problem.

6. What's changed since the last activity date

You are probably a newer model than whatever last worked here. Look at what changed between the last-activity date and now, and what you'd do differently:

  • Runtime and language features that replace hand-rolled code or a dependency (new Bun APIs, newer TypeScript or ECMAScript features).
  • Deprecated APIs, config options, and patterns still in use (tsconfig.json, bunfig.toml, workflows, scripts).
  • Better tools or services for jobs the project does in a roundabout way.
  • Playbooks: fetch https://raw.githubusercontent.com/rapideditor/agent-practices/main/playbooks/README.md and check each playbook's "Applies when" against this project. Make each one that applies its own item, linking the playbook, and note any order between them.
  • Agent instructions in AGENTS.md and .claude/skills/: workarounds for older models' limits that are no longer needed, instructions that host features now cover, and stale project facts. If the fix belongs in a synced file, say it should go to agent-practices rather than be changed locally.

Be specific: name the file, what to change, and why it's better now. If you aren't sure something is newer than the last-activity date, or stable yet, check the docs or say so. Don't pad this section; no items is a fine answer.

Other ecosystems

BunnpmCargoPython (uv)
Install from lockfilebun install --frozen-lockfilenpm cicargo build --lockeduv sync --locked
Auditbun auditnpm auditcargo audituvx pip-audit
Outdatedbun outdatednpm outdatedcargo outdateduv pip list --outdated

cargo audit and cargo outdated are separate installs. If one is missing, say so rather than installing it.

Report

Start with a short Where things stand paragraph: last-activity date, branch and working-tree state, CI on the default branch, and whether the baseline build and tests pass.

Then one numbered list (numbers run across sections, so the user can say "do 1, 3, and 5"), grouped by section: Baseline, Security, Dependencies, Sync, and Since {last-activity date}. Skip empty sections. Mark each item:

  • 🔥 Broken or risky — fix soon (failing build, exploitable vulnerability, red CI)
  • 👉 Worth doing (outdated dependency, deprecated option, sync behind)
  • 💡 Worth tracking (bigger change, a newer approach to consider)

End with what to do first, and whether /suggest, /bughunt, or /triage (for a backlog of open issues and PRs) is worth running next.

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.