How It Works
Open this page and you get a code of your own. Run the command on your server and the page follows each check live, then shows a scored report with what to fix first. The script comes straight from GitHub, so you can read every line before you run it.
Read-Only
It changes no setting, installs nothing and restarts nothing. It reads configuration, lists what is listening, asks the package manager what updates are waiting (without downloading anything) and reads the login log. With sudo it can read more (the firewall, password hashes, every user’s scheduled jobs, the login log); it still only reads.
What It Checks
Logins: whether root can log in over SSH with a password, whether SSH takes passwords at all, the SSH port and tries allowed, other accounts with root powers and accounts with no password. Open services: databases and admin services listening on every address, whether a firewall filters incoming traffic, and the Docker socket. Updates: waiting security updates, automatic updates, and whether the newest kernel is running. Signs of compromise: known crypto-miners, connections to mining-pool ports, scheduled jobs that download and run code, programs in temporary folders, and root’s SSH keys. Login attacks: failed logins in the last 24 hours, the addresses behind them (checked against Frabs’ threat feeds) and whether anything blocks them.
The Score
Every server starts at 100. Each problem takes off 15 and each warning 5. 90 and above is an A, 75 a B, 60 a C, 40 a D, and below that an F.
What It Sends
Each check’s result, the OS and kernel, and the addresses that failed to log in to your server, so they can be checked against threat lists. Never your server’s hostname or addresses, passwords, keys, file contents or command lines. Reports stay viewable at your link for 7 days.