agentleFS
jMertajMerta/codex-skills/vps-checkup/SKILL.md

vps-checkup

Skill130 starsChanged 11 months agoMITView on GitHub

Use when the user asks for a read-only health, security, network exposure, update, backup, or Docker audit of an Ubuntu VPS over SSH. Produces an evidence-based report and requires explicit approval before package refreshes, configuration changes, restarts, pruning, or reboot.

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are 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.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do not request or expose private keys, tokens, environment values, or container secrets.

    line 12
  2. Very specific:

    Check fail2ban only if installed and report missing tooling as not installed, not as failure.

    line 20
  3. Distinctive:

    If Docker exists, inspect daemon/rootless mode, containers, health/restarts, resource snapshot, disk use, logging growth, image age, and …

    line 22
  4. Very specific:

    Use references/ubuntu-docker-checkup-commands.md as a menu and run only commands relevant to the detected host.

    line 25
  5. Very specific:

    Use references/report-template.md.

    line 29

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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

VPS checkup

Confirm scope

Obtain the SSH target, required public services and ports, expected workloads, and whether passwordless sudo is available. Treat the audit as read-only. apt update changes package indexes and requires explicit approval even during a read-only checkup.

Do not request or expose private keys, tokens, environment values, or container secrets. Redact public output that includes credentials or sensitive hostnames.

Collect evidence

  1. Record host identity, UTC/local time, uptime, OS, kernel, virtualization, load, memory, disk space, inode use, mounts, and failed services.
  2. Inspect critical journal entries within a bounded time window; do not dump entire logs.
  3. Inspect effective SSH configuration and listening sockets.
  4. Detect the active firewall rather than assuming UFW: check UFW, nftables, then iptables as available. Compare exposure with the required port list.
  5. Check fail2ban only if installed and report missing tooling as not installed, not as failure.
  6. Inspect time synchronization, unattended upgrades, pending packages from current indexes, and reboot-required state. Label package results stale when apt update was not approved.
  7. If Docker exists, inspect daemon/rootless mode, containers, health/restarts, resource snapshot, disk use, logging growth, image age, and Compose projects. Skip Docker sections when absent.
  8. Record whether backups exist and their last successful evidence when discoverable; do not trigger a backup or restore test without approval.

Use references/ubuntu-docker-checkup-commands.md as a menu and run only commands relevant to the detected host.

Report and changes

Use references/report-template.md. Rank findings by impact and likelihood, cite command evidence, distinguish confirmed findings from unknowns, and list the smallest recommended actions.

Obtain explicit approval for each mutating group, including package-index refresh, upgrades, configuration edits, firewall or SSH changes, service/container restarts, pruning, log deletion, reboot, and backup/restore operations. Before SSH or firewall changes, verify a second access path and configuration syntax. Recheck affected services and access after any approved change.

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.

Reviews can't be read right now.

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.