Skip to content

Security: Q29F/Next.js-de

Security

docs/security.md

Security boundary

The mathematical calculation runs in browser memory on fixed numeric operations. It does not evaluate supplied expressions, upload vector inputs, store them in a database, or invoke an external model or compute service. JSON downloads contain the calculation the user explicitly chooses to save.

Input handling

The mathematical module validates types, equal vector dimensions, component magnitudes, structural potentials, and finite outputs. Text inputs have length limits and a decimal-only parser. React renders values as text; application data is not inserted as HTML. The vector size limit bounds plotting and calculation work.

Browser policy

proxy.ts generates a fresh cryptographic nonce for each page request and supplies it to both Next.js rendering and the response's content security policy. Dynamic rendering avoids caching HTML with a reused nonce. Production policy permits the application's required scripts and styles, blocks inline evaluation, frames, object content, base-URL changes, and form submissions, and restricts connections to the same origin. Development-only exceptions support the framework debugger and hot reload.

Responses also apply nosniff, frame denial, a no-referrer policy, and disabled camera, microphone, geolocation, and payment permissions. The framework branding header is disabled. The liveness endpoint is read-only and does not disclose credentials, environment values, or internal paths.

Verification and operations

Production tests check response headers, fresh nonces, script markup, unsupported HTTP methods, and browser operation with CSP enforced. Numerical tests cover the fixed mathematical boundary. Dependency audits run against the committed lock.

The hosting environment is responsible for TLS, network access policy, resource allocation, logs, and platform updates. This application does not supply multiuser authentication or claim to assess the security of its hosting infrastructure.

There aren't any published security advisories