NOVA Verify & Security Scan

Docs › NOVA MCP

Verify & security

NOVA Verify checks a project before it goes live. Your AI tool runs it with verify; every deploy also runs the security scan first.

A security scan, then the app itself, then the verdict: only a live secret or a critical vulnerability blocks a deploy.

The security scan

SecretsLive keys and tokens written in the code (Stripe, AWS, GitHub, AI providers, private keys, database URLs with a password). Shown masked, never in full.
DependenciesExact versions from your lockfile checked against OSV, the open vulnerability database. Version ranges with no lockfile are not guessed.
CodePatterns like SQL built from strings, running code from a request, debug mode in production, HTTPS checks turned off.
SettingsPrivate files in the project, a .env file a static site would serve, admin routes with no sign-in check.

The app itself

A websiteThe same build check and real-browser test as Make Live.
A server appNOVA starts it in a sandbox and watches what it prints for 45 seconds, to catch what only breaks at runtime: a missing environment variable, a crash on the first request.

What blocks a deploy

Only two things: a live secret in the code, or a dependency with a critical known vulnerability. Everything else is reported with how to fix it, and you decide.

Join our Discord

Chat with our engineers and become part of the active NOVA developer community.

Join Discord

Need help?

Open an issue on GitHub or email support@bridgeye.com.

Get help
Bridgeye • Autonomous Software Engineering.