Guide

You Can Secure It Yourself. Should You?

The topic

You've built a habit-tracking app with a hosted database behind it. It has accounts, it stores each user's entries, and it works beautifully. Every screen shows the right person the right data.

What you haven't checked is what happens if someone skips the screens entirely. Open the browser's developer tools, copy the database key from the page's JavaScript, and query the database directly. If the access rules were never switched on, that one key can read every user's data. This exact pattern turns up again and again in apps built quickly with AI help, because the app looks finished from the front.

You can secure it yourself. Should you?

The discussion

Can you?

Mostly, yes. The common holes are well known and fixable by any developer who goes looking:

  • Database access rules (row-level security or equivalent) switched off, or written so loosely they allow everything.
  • Secret keys shipped to the browser in a front-end bundle or committed to a public repository.
  • API routes that check you're logged in but not whether the record you're asking for is yours.
  • No rate limiting on login or password reset, so guessing is free.
  • Dependencies with known vulnerabilities that nobody has updated.

An AI assistant is genuinely useful here too. Ask it to review your code as if it were trying to break in and it will find some of these.

Is it cheaper?

A self-review costs a day or two. An independent review costs money. Both are trivial next to the cost of a breach, which for a UK product handling personal data includes the obligation to report it to the ICO within 72 hours where there's a real risk to people (UK GDPR; covered in our GDPR guide), telling affected users, and the trust you lose with every one of them.

Is it better?

This is where doing it yourself has a real weakness. You are reviewing your own assumptions. The same mental model that built the app reviews the app, and an AI assistant prompted by you inherits your blind spots too, because it only looks where you point it.

Fresh eyes find different things. That's the whole reason code review exists in every serious engineering team, and security is where it pays off most.

So, should you?

Do the self-review first, always. Work through the list above on your own app this week. It's the cheapest security you'll ever buy.

Then, before you take payment or store anything sensitive, get someone who didn't build it to look. That doesn't have to be an expensive formal audit for a small product. It does have to be someone other than you and the assistant you built it with.

Where we can help

Compliance review is part of the production groundwork in LaunchUK, and our guide to GDPR for a one-person SaaS covers what the law expects of you if something does go wrong. We don't offer penetration testing or certified audits; what we can give you is a second pair of eyes on how your app is put together. Get in touch if you'd like one.