| name | linux-server-hardening |
|---|---|
| description | Make a Linux box survive unattended — panic auto-recovery, crash forensics that actually work, boot paths that cannot stall, and how to prove each recovery path with a real fault injection. Use this whenever a machine "went offline", "hung", "froze", "died overnight", needed a manual power cycle, or whenever setting up a mini PC, NUC, homelab node, or any host nobody is sitting in front of. Also use it before enabling kdump, before trusting a watchdog, or when someone asks why a server did not reboot itself after crashing. |
A server that nobody watches fails differently from a desktop. The distro defaults are tuned for a human at the console, and several of them turn a 30-second incident into a multi-day outage. This skill is the checklist, the