
FRODO Security Watchdog | Sicherheit für Linux-Server
At Estudio Ideas, we manage the hosting infrastructure of over 300 WordPress sites, all running on our own server, Frodo. With that scale, one question became inevitable: how do we detect a threat on a compromised site before it becomes a problem for the client — and how do we do it without risking deleting something we shouldn't?
The answer was to develop our own internal tool: FRODO Security Watchdog.
Das Problem mit herkömmlichen Malware-Scannern
Most security tools for WordPress do one of two things: they alert and let a human decide, or they act automatically and — in the worst case — delete files. The second option is quick, but it is also the most dangerous: a false positive can delete a legitimate plugin, a critical configuration file, or evidence that is later needed to understand what really happened.
Wir haben einen anderen Weg gewählt.
Die goldene Regel: Es wird nie etwas gelöscht
FRODO Security Watchdog is built on a non-negotiable principle: no file is ever deleted, under any circumstances. When the system detects a critical threat — a known webshell, obfuscated malicious code, a backdoor — it does not delete it. It moves it to an isolated quarantine zone, but first:
- Calculates its digital fingerprint (SHA256) to verify that the copy is identical to the original.
- Erst wenn diese Kopie geprüft ist, entfernen Sie die Datei von der Live-Website.
- Generates a complete log: date, time, account, domain, original path, and reason for detection.
If it ever turns out to be a false positive, restoring the file is trivial — everything is tracked and available.
Eine Risiko-Engine, kein Alles-oder-nichts-Schalter
Not all findings are the same. A loose PHP file in an uploads folder is not the same as aeval() with obfuscated code in Base64. FRODO Security Watchdog assigns a score to each finding and classifies them:
| Stufe | Bereich |
|---|---|
| 🟢 Grün | 0–20 |
| 🟡 Gelb | 21–50 |
| 🟠 Orange | 51–80 |
| 🔴 Rot (kritisch) | 81+ |
Only individual findings that reach the critical level trigger an automatic quarantine action. Everything else is documented for human review — preventing the system from "overreacting" to low-risk signals.
Forensische Fallakten, nicht nur ein Log
When a real threat is detected and quarantined, the system automatically generates aFallakte mit einer eindeutigen ID (FSW-2026-000124, for example): hash SHA256 and MD5 of the file, permissions, owner, the signature that detected it, the risk level, and a complete timeline. That document can be directly attached to an audit or incident response report — the kind of evidence a client needs when they ask "what exactly happened and when?"
Für echte Größenordnungen gebaut, nicht für eine Demo
One detail that is not often mentioned: the first version of our detection engine, when tested against a simulated catalog of 300 accounts with tens of thousands of files, took so long that it didn't even finish in a reasonable time. The bottleneck was technical and subtle (thousands of unnecessary external processes launching), and we resolved it until we achieved that a complete run on that same scale takes minutes, not hours. That kind of optimization work is not visible, but it is what differentiates a tool that works on paper from one that works in production, every day, twice a day.
Entwickelt, um mit unseren übrigen Sicherheitsmaßnahmen zusammenzuarbeiten
FRODO Security Watchdog does not replace the rest of our security infrastructure — it complements it. It works alongside Cloudflare, BitNinja, and CSF/LFD, respects a configurable whitelist to not interfere with legitimate integrations of each client, and runs with low CPU and disk priority to not affect the performance of active sites while scanning.
Was das für unsere Kunden bedeutet
In practice: twice a day, each site under our management is checked for webshells, suspicious changes in core WordPress files, unauthorized new admin users, plugins with known historical vulnerabilities, and more than a dozen other signs of compromise — all without manual intervention, and without the risk of aggressive automation causing more harm than the threat itself.
It's the kind of infrastructure work that usually goes unseen, but makes the difference when something does happen.
Do you want to know more about how we protect the sites we manage, or do you need help with the security of yours? Let's talk.





Allgemeine Themen
Neue Services
Support
Support











































