A simple upload button sits on your website, letting customers submit a profile photo or a document. It looks harmless enough. To an attacker, it can be an open invitation to place their own code directly onto your server and run it whenever they choose, sometimes for months without anyone else knowing it is there.
Why upload features are riskier than they look
File upload functionality exists on countless business websites, from CV submissions on a careers page to document sharing on a client portal. Each one asks the same underlying question: what exactly is this system willing to accept, and what happens to it afterwards once it lands on the server? If a site checks only the file extension, an attacker can rename a malicious script to look like an innocent image and slip straight past that check without much effort at all, since the file name alone tells the server nothing reliable about the content inside it. If the server then executes files from the upload directory, that disguised script becomes fully functional code running with the server’s own permissions.
This is how a web shell gets planted, a small script that gives an attacker a persistent, remote command line into your infrastructure, often surviving quietly for months before anyone notices unusual activity in the logs. Thorough web application pen testing specifically probes upload functionality with disguised file types and unexpected content, because this is exactly the kind of flaw that automated scanners often miss entirely, relying instead on a human tester who knows what to try.

Small oversight, complete compromise
What makes insecure uploads so serious is the scale of what follows. Once a web shell is in place, an attacker typically has the same level of access as your web application itself, which often means reaching the database, other files on the server, and sometimes a path into the wider internal network if segmentation is weak or non-existent. A single missing check on file type or execution permissions can be the entire difference between a contained issue and a full breach affecting every customer record you hold.
William Fieldhouse has traced entire compromises back to exactly this kind of gap.
“On a recent web application test, I uploaded a working script disguised as a profile picture, and it gave me a command line into the client’s server within about ninety seconds of finding the flaw”
— William Fieldhouse, Director of Aardwolf Security Ltd
Ninety seconds is not an exaggeration for effect. It reflects how little stands between a poorly configured upload feature and complete server access once an attacker knows precisely what to try, and a properly built API sitting behind that same upload form deserves exactly the same scrutiny as the front end, since the underlying weakness is often identical.
Treat every upload as untrusted until proven otherwise
Fixing this properly means validating file content rather than trusting extensions, storing uploads outside any directory capable of executing code, and scanning everything that comes in before it touches your live environment. These are established, well-documented fixes, but they only work if someone actually checks whether they have been applied correctly across every part of the site that accepts a file. Get in touch with Aardwolf Security for API pen testing that tests your upload functionality the way a real attacker actually would.
