Malware scanners for Linux servers
Most malware on Linux servers is not a virus that arrives by email. It is a cryptominer dropped through a vulnerable web app, a web shell in an upload folder, or a backdoor added after a stolen SSH key. A scanner finds the files it has signatures for; this guide covers the four open-source ones you are most likely to meet, and what none of them can see.
| Scanner | What it looks for | How it runs |
|---|---|---|
| ClamAV | Files matching its signature databases: viruses, trojans, malware in archives and documents, ELF binaries | Command-line scanner or a daemon; freshclam updates signatures; on-access scanning on Linux through ClamOnAcc. GPLv2, by Cisco |
| Linux Malware Detect (LMD, maldet) | Malware common on shared web hosting, by MD5/SHA-256 hashes, hex patterns, YARA rules and statistical checks; can use ClamAV as an engine | Scheduled scans of web roots, or real-time monitoring with inotify; quarantine and restore. GPLv2, by R-fx Networks |
| rkhunter | Known rootkits, changed system binaries and suspicious settings | Scheduled scans compared against a stored baseline; last release 1.4.6, February 2018 |
| chkrootkit | Signs of known rootkits, worms and kernel-module trojans; deleted login records | A shell script and small helpers you run by hand or from cron; 0.60 released October 2026 |
Which one to run
- A web server or shared hosting: LMD on the web roots, with ClamAV as its engine, because web shells and injected PHP are what it was written for.
- Any server that stores files from users or email: ClamAV, with signatures kept up to date.
- A rootkit check after something looks wrong: rkhunter or chkrootkit, then a second opinion from a clean boot medium, because a rootkit can hide from tools running on the system it controls.
What a scheduled scan misses
Signature scanners find files someone has already seen and described. They do not see malware that only runs in memory, a new variant nobody has written a signature for, or an attacker using the tools already on the server. A scan every six hours also leaves six hours in which a miner can run before anyone looks.
Watch behaviour as well as files. A web server starting a shell, an unknown process using all the CPU, a new outbound connection or a new cron entry are the signs that something got in, whatever the files look like.
Mind the memory. ClamAV loads its whole signature database into memory, which is a real cost on a 1 GB VPS. Check that the scan actually ran, not just that it is scheduled.
Where Vigilhold fits
Vigilhold is a hosted security agent for Linux servers and Docker hosts. Its installer adds ClamAV, and rkhunter where the distribution packages it, runs their scans on a schedule at low priority, skips a ClamAV scan when the server lacks the memory, and shows the results in the dashboard next to everything else it watches: processes, network connections, file changes and logins. It responds only within limits you set, and every action is in an audit log you can undo from.
Checked in November 2026: ClamAV documentation, ClamAV on-access scanning, Linux Malware Detect on GitHub, Rootkit Hunter, chkrootkit.
Vigilhold vs fail2ban · Vigilhold vs CrowdSec · Vigilhold vs Wazuh · Start a 14-day free trial