In short
A website or app works, so it can go live. But does it also work securely? When we check small sites before go-live, we find the same weak spots time after time. This is the list we work through ourselves.
The checklist
- Admin functions properly shielded. Is every function that changes something (save, delete, upload) secured on the server, and not just hidden in the menu?
- No secrets in the code. Passwords, API keys and default keys do not belong in the code or in git, but in the server's configuration. If a default key is still active, replace it.
- Never trust input. Text that users enter is displayed safely (no injected scripts) and cannot manipulate an email through a contact form.
- Encrypted connections, properly verified. HTTPS everywhere, and nowhere a certificate check that was once switched off "just for testing".
- Limit attempts. Logins and forms have a limit, so nobody can try passwords or send spam endlessly.
- Security headers. Settings that tell the browser what is and is not allowed, such as blocking the page from being loaded inside someone else's page.
- Backups that work. Automatic, stored away from the server, and restored at least once as a test.
- Personal data protected. Only what is needed, properly shielded, and clear about who has access.
- Error messages without details. The user sees a tidy message; the technical details are only in the log.
- Automated tests. So that a later change does not quietly undo a fix.
What to watch out for
Security is not a one-off job. Record the checklist in the project and work through it again with every major change. In AI-assisted development this belongs in the house rules: whatever AI builds is always reviewed on these points by a senior developer.
Would you like advice on the security of your software? Request a free advice conversation.
Updated on