Security Checklist Before Shipping a Game Built with Cursor, Claude or Lovable
AI coding tools help you ship fast, but they optimize for 'it works', not 'it's safe'. Run through this checklist before you hit publish.
1. Secrets and API keys
- Search the project and git history for keys: Steam, Stripe, PayPal, OpenAI/Anthropic, Firebase Admin, Supabase service_role, Discord bot tokens, cloud credentials.
- Move every secret into server-side environment variables. Nothing secret in client code, config files or build output.
- Check that
.envis in.gitignore, and that you're not exposing secrets throughVITE_/NEXT_PUBLIC_prefixes. - Rotate any key that was ever committed or shipped. Deleting it in a new commit isn't enough.
- Scan with a tool like
gitleaksor GitHub secret scanning.
# Scan the repo and its full history for leaked secrets
gitleaks git -v .
2. AI API usage
- If your game calls an LLM, route it through your server so the key stays hidden.
- Add per-user rate limits and a spending cap on the provider dashboard so one person can't run up your bill.
- Treat AI output that drives game logic as untrusted input; players will try prompt injection.
3. Database and backend
- Firebase: no test-mode rules left; default deny; players can only touch their own documents.
- Supabase: RLS enabled on every table in exposed schemas; check the Security Advisor.
- Admin and debug endpoints are removed or behind authentication.
- Storage buckets aren't publicly writable or listable unless intended.
Details: Firebase and Supabase security rules for multiplayer games.
4. Multiplayer and cheating
- The server decides damage, movement limits, loot, currency and scores.
- Every incoming message is validated for type, range, ownership and rate.
- Leaderboard writes go through a server function, never straight from the client.
Details: Never trust the client.
5. Authentication and platform
- Verify Steam players with auth tickets on your server, not by trusting a SteamID from the client. See keeping your Steam Web API key safe.
- Session tokens expire, and password reset and email flows can't be abused to take over accounts.
6. Payments and in-game purchases
- Grant items only after your server confirms payment through the provider's webhook (for example a Stripe webhook with signature verification) or platform receipt validation.
- Never trust a client message that says "purchase succeeded".
- Prices come from the server, not the client.
7. Dependencies and plugins
- Run
npm audit(or your ecosystem's equivalent) and update packages with known vulnerabilities. - Check that every package your AI tool added actually exists and is the real one. AI assistants sometimes suggest package names that don't exist, and attackers register look-alike names.
- Only use asset store plugins and mods from sources you trust.
8. Servers and hosting
- Only the game port is open to the internet. Databases, Redis and admin panels are not.
- Default passwords changed; SSH uses keys, not passwords.
- HTTPS/WSS for all web traffic; CORS limited to your own domains.
- Debug builds, source maps with secrets, and verbose error pages are off in production.
9. Player data and privacy
- Collect only what you need. Know where it's stored and who can access it.
- Have a simple privacy policy, and check the rules that apply where your players live (for example GDPR in the EU or COPPA if you target kids in the US).
10. One last review
Ask your AI tool to review its own work as an attacker would, then verify the risky parts yourself:
Act as a security reviewer. List every way a player could cheat, steal data,
leak secrets or run up my costs in this project, ranked by severity, with the
file and line for each and a concrete fix.
Want this checked automatically?
BossCheck is a security scanner for indie games built with AI tools. It's in pre-launch: join the waitlist for early access.
Join the waitlist →FAQ
Are games built with AI coding tools less secure?
Not automatically, but AI tools tend to pick the quickest working solution, which often means hard-coded keys, open database rules or trusting the client. A security pass before launch catches most of these.
What's the most common security mistake in vibe-coded games?
Common ones include secrets in client code or git history, database rules left open, and multiplayer servers that accept whatever the client sends.
Do single-player games need a security check?
Yes, if they use any online service: leaderboards, cloud saves, payments, AI APIs or analytics keys can all be abused.
How long does this checklist take?
For a small project, a careful pass usually takes an afternoon. BossCheck is being built to automate it in minutes.
More guides
- How to Keep Your Steam Web API Key Out of Your Game Build
- Firebase and Supabase Security Rules for Multiplayer Games
- Never Trust the Client: Stopping Cheaters in Vibe-Coded Multiplayer Games
- Is Code from Cursor, Lovable or Claude Secure? What to Check Before You Ship
- How to Protect API Keys in a Unity Game
- How to Stop Players Editing Browser Game Saves to Unlock Paid Items
- Securing Godot 4 Multiplayer: RPCs, Authority and Server-Side Validation
- Is My Vibe-Coded Game Safe? A Free 10-Minute Security Self-Check
- Stop Bots and Scripts from Breaking Your Game Economy Before a Steam Launch