Web Application Security Checklist: 12 Things to Do Before Launch
Your site receives hundreds of requests daily from scanner bots, automated attacks and manual probing — from day one. Most breaches are not cinematic: one SQL query without parameters, an admin panel with a weak password, or an outdated package with a known vulnerability. This is the checklist we run at GeeDook before every launch.
1. Enforce HTTPS with HSTS
An SSL certificate alone is not enough. Set the Strict-Transport-Security header so browsers always connect over HTTPS. The HTTP-to-HTTPS redirect must happen server-side, not just client-side.
2. Parameterize every database query
SQL injection is still among the most common attacks. Anywhere user input lands directly inside a query, the door is open. An ORM with bound parameters (like Entity Framework) solves this by default — as long as you don't write raw queries.
3. Encode output to prevent XSS
A script that enters the page from user input (a comment, a profile, a form) executes on every view. Every piece of user content must be encoded before rendering — or if raw HTML is allowed (like an admin editor), it must come from admins only.
4. Build real authentication
Passwords hashed with a standard algorithm (bcrypt/Argon2), account lockout after failed attempts, and rate limiting on the login endpoint. For admin panels, take two-factor authentication seriously — an 8-character password alone is not security.
5. Treat tokens like money
JWT in localStorage is an XSS away from compromise. Short token lifetimes, refresh with rotation, and server-side revocation. Admin tokens must never appear in URLs or logs.
6. Keep dependencies updated
A large share of breaches come from outdated packages with known vulnerabilities. Put an automated dependency scan in your CI — it costs nothing and reports weekly.
7. Move secrets out of code
A database password or API key inside source code means anyone with repo access owns your infrastructure. Secrets belong in environment variables or a secrets manager; even better, limit blast radius so one leak is not a catastrophe.
8. Backups without restore drills are not backups
Take daily backups, but actually rehearse a restore every month. Backups that live only on the same server die with the server — one copy must be off-site.
9. Apply least privilege
The application's database account should not have DROP TABLE rights; server admins should not do daily work as root. Every service gets only the access it needs.
10. Set up logging and monitoring
A breach nobody sees runs for months. Log logins and logouts, log 500 errors, and alert on suspicious patterns (repeated failed logins). Have uptime monitoring — a down site must not go unnoticed.
11. Add a WAF and rate limiting
A WAF (even Cloudflare's free tier) catches the first wave of attacks; rate limiting on forms and APIs stops spam and brute force.
12. Get an independent audit before launch
A developer doesn't see the things they wrote themselves. An independent security audit before launch is the cheapest insurance you will ever buy.
---
Frequently asked questions
How do I secure a WordPress site?
WordPress itself is fine; plugins usually aren't. Auto-updates, removing inactive plugins and a WAF are the minimum. For commercial sites, an annual audit is recommended.
How long does a security audit take?
For a mid-sized web application, one week at GeeDook: configuration review, dependencies, authentication, and a prioritized report with remediation estimates.
---
Does your system pass this checklist today? Order a fixed-price security audit — you keep the full prioritized report even if you never ask us to do the fixes.
The security checklist for shipping a public API