Vibe Coding Security Checklist: What to Check Before You Launch
Check your own app — free, in about a minute.
The free mini-scan checks your live homepage for TLS/HTTPS basics, response security headers, cookie flags, exposed sensitive files, Firebase/Supabase backend rules, and the WordPress plugin versions your site exposes. You get a score and the finding titles — no account, no card.
Run the free mini-scan →Vibe Coding Security Checklist: What to Check Before You Launch
You built it with AI. It works. Now you're about to put it on the internet where strangers can poke at it.
This checklist covers what to verify before launch. It's organized by area, ordered roughly by how often each issue appears in AI-generated apps. You don't need to be a security engineer — you just need to be methodical. Work through it top to bottom, and don't ship until every box is checked or consciously accepted.
Authentication & Sessions
- [ ] Every protected route enforces auth on the backend, not just the frontend. Hiding a page behind a login screen means nothing if the API serves data to unauthenticated requests. Test with curl while logged out.
- [ ] Passwords are hashed with a proper algorithm (bcrypt, scrypt, or argon2). Search your code for MD5, SHA-1, or — worse — plaintext password storage.
- [ ] Session tokens expire. Tokens that live forever are tokens waiting to be stolen. Check for reasonable expiry and a logout flow that actually invalidates the session server-side.
- [ ] Password reset can't be abused. The reset flow should use single-use, expiring tokens sent to the account's email — not guessable URLs, not security questions.
- [ ] Failed logins are throttled. After a handful of bad attempts, slow down or lock out. Otherwise you're handing attackers a credential-stuffing target.
Data & Database
- [ ] No raw SQL built from user input. Search for string-concatenated queries. Use parameterized queries or your ORM everywhere, including search filters and admin tools.
- [ ] Users can only access their own data. For every endpoint that takes an ID, verify the current user owns that record. Test with two accounts and swapped IDs (this catches IDOR).
- [ ] Sensitive fields are never over-exposed. API responses shouldn't include password hashes, internal IDs, or other users' emails "just in case the frontend needs them." Return the minimum each endpoint needs.
- [ ] File uploads are restricted. If users can upload files, validate type (by content, not extension), limit size, store outside the web root, and never execute what was uploaded.
Secrets & Configuration
- [ ] No API keys, tokens, or credentials in client-side code. Search your bundled JS for
sk_live, api_key, secret, AKIA, and token. Anything found there is public.
- [ ] No secrets committed to your repo. Check git history too, not just the current files — keys live on in old commits. If a key was ever committed, rotate it.
- [ ] Production uses environment variables, not hardcoded config. Database URLs, API keys, and feature flags should come from the environment.
- [ ] Debug mode is off in production. No stack traces to users, no debug routes (
/debug, /test), no verbose logging of sensitive data.
Frontend & User Input
- [ ] All user input is escaped on output. Anywhere your app renders user-supplied text — comments, names, messages — verify it can't inject HTML or JavaScript. Test with
<img src=x onerror=alert('xss')>.
- [ ] Forms have CSRF protection (if your framework doesn't handle it automatically). State-changing requests shouldn't be triggerable from a malicious third-party page.
- [ ] Security headers are set. At minimum:
Content-Security-Policy, X-Content-Type-Options: nosniff, X-Frame-Options (or frame-ancestors), and Strict-Transport-Security. These are one-line wins in most frameworks.
- [ ] HTTPS everywhere, no mixed content. Your site should redirect HTTP to HTTPS, and every asset (scripts, images, API calls) should load over HTTPS.
APIs & Abuse Prevention
- [ ] Rate limiting is on. Login, registration, password reset, contact forms, and public API endpoints should all throttle aggressive clients. Hammer each one with a quick script and confirm you get 429s.
- [ ] CORS is locked down. Your API shouldn't accept cross-origin requests from
* unless it's genuinely a public API. Restrict to your own domains.
- [ ] Error messages don't leak internals. "Invalid password" vs "user not found" tells an attacker which accounts exist. Keep auth errors generic. Stack traces stay server-side.
- [ ] Admin functionality is actually restricted. Check every admin route and function with a non-admin account. "Nobody knows the URL" is not access control.
Infrastructure & Launch Hygiene
- [ ] Dependencies are up to date. Run your package manager's audit (
npm audit, pip audit, etc.) and fix critical vulnerabilities before launch.
- [ ] Backups exist and you've tested a restore. A backup you've never restored is a hope, not a plan.
- [ ] You have logging for security events. At minimum: logins, failed logins, password changes, and admin actions — somewhere you can actually review them.
- [ ] A contact path exists for security reports. Even a simple
security@yourdomain.com (or a note in your footer) means researchers tell you instead of tweeting about it.
How to Work Through This
Be honest about the ones you skip. Every unchecked box is a risk you're accepting — some are fine for a weekend project, none are fine for an app handling other people's data or money.
Two practical tips:
- Test like an attacker, not like a user. Users click the happy path. Attackers try the logged-out curl request, the swapped ID, the weird input. For each check above, ask "what would someone trying to break this do?"
- Re-check after every big AI-assisted change. Vibe-coded apps evolve fast. A refactor that "just cleaned things up" can quietly drop an auth check. Run the checklist again before each significant deploy, not just the first launch.
Don't Want to Do It by Hand?
Working through 25 checks takes real time — and it's easy to miss the one that matters. PreShipIt does the attacker-style probing for you: paste your app's URL, and you get a plain-English report of what to fix before you ship.
$29, one-time. No subscription. Get your pre-ship inspection at preshipit.com
The PreShipIt Brief — a short weekly email for people shipping vibe-coded apps: the week's important vulnerabilities and launch-safety checks, in plain English.
New and free. Double opt-in: we email you a confirmation link before anything else arrives. Unsubscribe any time.
Check your own app — free, in about a minute.
The free mini-scan checks your live homepage for TLS/HTTPS basics, response security headers, cookie flags, exposed sensitive files, Firebase/Supabase backend rules, and the WordPress plugin versions your site exposes. You get a score and the finding titles — no account, no card.
Run the free mini-scan →
← Back home · Run a free mini-scan · Scan my app — $29 · See a sample report