About AGuard
Website security,
from design down to the kernel.
18 years on the web: design, SEO, development, servers. When you have built websites yourself, security stops being abstract. You know the exact places that usually go unchecked.
2008–
Web design and SEO
~2012
Programming and development
~2015
Server administration
Today
Website security
2008
Started working on the web
Design, SEO, development, security
18
Years of web stack experience
The whole stack, not only security
13 000+
Websites in a single incident
One of the most complex cases
0
Clients named publicly
An incident is a private matter, not a portfolio item
Who we are
Not only security specialists.
Web developers who understand an attack from the inside.
In 2008 we worked on design and SEO. Development, programming and server administration came later. We got into security through practice: when you have built a website yourself, you know where access is typically left open.
Most of the people who contact us are company managers and those responsible for IT whose websites were infected. They do not know whether the hosting provider cleaned everything or only the surface. We work directly with the client, with no intermediaries: one person responsible, and a specific report when the work is done.
We are based in Vilnius and work with clients across Europe. All work is done remotely. Where your server is physically located does not matter.
From practice
13,000 websites. Two weeks of cleanup.
One open port.
A hosting server in the Czech Republic compromised by outsiders: several unrelated mistakes at the same time, none of which looked critical on its own. The chain below shows how they connected. We were called in to clean up the result.
01 — Initial entry point
An outdated WordPress install with a plugin that had not been updated.
02 — Database credentials
Port 3306 open to outside connections, plus a backup copy wp-config.php.back in the public directory: the login details were readable without authentication.
03 — Administrator account and backdoor
An administrator user was created directly in the database. A backdoor file was then uploaded through the WordPress admin.
04 — Privilege escalation to root
A copy-on-write kernel vulnerability (the Linux kernel had not been updated) took the attacker from a website user to server administrator.
05 — Mass compromise of every website
Access to all 13,000+ websites on the server. A backdoor that injects redirect code was uploaded to each of them.
06 — Result: casino redirects. Cleanup took two weeks.
Every website sent visitors arriving from search engines to an illegal casino. Search rankings collapsed, reputations were damaged and the hosting provider imposed penalties.
The lesson: every insecure component (an open port, an outdated kernel, a publicly accessible backup file) is a potential bridge to the whole infrastructure. Especially when one server hosts the websites of several clients and those websites store customer data.
Complex cases
When the incident is not a standard one
Several websites infected at once, shared hosting with dozens of clients, a deep break-in with chains of backdoors: cases like these need a systematic approach, not a quick fix.
We decide what to do after a full analysis: file system, database, cron, server configuration, logs. Only then come cleanup and prevention.
Method
Antivirus software
looks for signatures.
We look for the logic.
Automated scanning detects known variants of malicious code. An experienced attacker takes advantage of that and leaves access where the scanner does not look. We check the logic: whether the access can restore itself, what has changed in the list of administrators in the database, whether .htaccess rewrites browser requests.
SSH access that works without a password
Key pairs left behind by an earlier break-in
Cron jobs that restore the backdoor file
Every few minutes, even after a cleanup
PHP code in JPEG or PNG metadata
Scanners check file extensions, not file contents
Malicious fragments inside legitimate plugins
One line among 4,000 lines of real code
Changes to the CMS core code
Core files are the last place scanners look
Open access left over from an old infection
The hosting provider cleaned up, but not completely
How we work
Principles we do not change
A clear price before work starts
We assess the situation, name the price and only then start. No surprises on the invoice.
Credentials only through a secure channel
We accept passwords only through a secure channel agreed before work starts: our own encrypted form or SSH keys.
A report after every job
What was found, what was removed and what was changed: all of it is documented.
Client confidentiality
We do not publish client names. A security incident is a private matter for your business, not marketing material for us.
Specialisation
We know our field. And its limits.
Our field
We work with
- PHP / MySQL websites and online shops
- Node.js and modern web applications
- Any hosting provider, in Europe or elsewhere
- Cloudflare and WAF configuration and monitoring
Not our field
We will tell you if it is outside our field
- Penetration testing and red team operations
- GDPR and legal compliance consulting
- Ransomware negotiations
- Corporate IT security (Active Directory, SOC)
Business details
Get started
Have a question or an unusual situation?
Describe what is happening and we will reply within one business day.



