a1mb0t38/medicare-connect

Medicare Connect presents a useful full-stack healthcare workflow with distinct patient, doctor, and administrator experiences. The Next.js App Router frontend includes doctor discovery, appointment scheduling, prescriptions, reviews, Stripe Checkout, Google and password authentication through Better Auth, and dashboard charts. The code is divided into reusable pages and components, dependencies are locked, and the linked Express and MongoDB API verifies JWT signatures before most data operations. Server-side session and token retrieval are also used on several protected pages. This is a meaningful feature set that demonstrates an understanding of authentication, payments, database-backed workflows, and role-oriented interface design.

The most important improvement is authorization. The API verifies that a token is valid, but it does not attach the token identity to requests, check roles, or confirm ownership of records. As written, any authenticated account could potentially list or delete users, verify doctors, change another appointment, create prescriptions, or request medical records by supplying another user ID. The frontend proxy protects only the doctor-details route, while several dashboard requests omit authorization headers. These controls should be enforced centrally on the server with explicit patient, doctor, and administrator policies, ownership checks, validated input, and audit logs.

The payment flow also needs stronger server-side guarantees. The checkout endpoint accepts the consultation fee and appointment details from the browser, so those values can be altered. Appointment creation occurs from the success page instead of a verified Stripe webhook, which creates replay and reliability risks. The server should calculate prices from trusted doctor records, process signed webhooks idempotently, and store the Stripe session or payment-intent ID. Automated tests and CI are absent, despite the sensitivity of authentication, payments, and prescriptions. Documentation should also be synchronized with the code: it references Next.js 15 and Firebase variables, uses a different clone name, contains placeholder screenshots, and claims MIT licensing without including a license file. Privacy, data-retention, security-reporting, and healthcare-data handling documentation would make the project much more credible.