PromptVault

A Spring Boot web application for COMP47910 – Secure Software Engineering, Assignment 1.

PromptVault is a prompt management system where users create, store, organise, and submit prompts to a simulated AI assistant. Admins manage users, prompt categories, policy keywords, and review flagged prompts.

Note: This is the Phase 1 ("Build it") version. Security hardening is intentionally minimal and is addressed in later assignments.

Tech stack

• Java 21 (compiles and runs on JDK 17+/21+/25)
• Spring Boot 3.5.16 (Spring Web, Spring Data JPA, Thymeleaf, Spring Security, Validation)
• MySQL 8 (run via Docker)
• Maven (via the included Maven Wrapper – no global Maven install needed)

Prerequisites

• A JDK (17 or newer; developed on JDK 25)
• Docker Desktop (to run MySQL), or a local MySQL 8 server

1. Start MySQL

Using Docker (recommended):

docker run --name promptvault-mysql -e MYSQL_ROOT_PASSWORD=root -e MYSQL_DATABASE=promptvault -p 3306:3306 -d mysql:8

This creates a database named promptvault with user root / password root, matching src/main/resources/application.properties.

If you use a local MySQL instead, create a database called promptvault and update the username/password in application.properties if needed.

2. Run the application

From the PromptVault folder:

# Windows
./mvnw.cmd spring-boot:run

# macOS / Linux
./mvnw spring-boot:run

Then open: http://localhost:8080

On first run the schema is created automatically by Hibernate, and sample data is seeded (see below).

3. Seeded accounts and data

Seeded automatically on first startup (only if the database is empty):

Role	Username	Password
Admin	admin	admin123
User	alice	alice123
User	bob	bob123

Also seeded: 3 categories (Coding, Research, Cybersecurity), 5 policy keywords (password, API key, secret, credit card, confidential), and 5 prompts (a mix of private and shared, two pre-flagged).

4. Features

Admin

• Login / logout
• Manage users (view list, enable / disable accounts)
• Manage prompt categories (add / edit / delete)
• Manage policy keywords (add / edit / delete)
• View flagged prompts (title, owner, category, keyword, date)

User

• Register, login / logout
• Create / edit / delete own prompts (title, text, category, visibility private or shared)
• View own prompts; browse other users' shared prompts only
• Submit a prompt to the simulated AI assistant
• Receive a warning and have the prompt flagged when it contains a policy keyword
• View own submission history

5. Database script

The schema is generated by Hibernate from the JPA entities, and sample data is inserted by DataSeeder. A standalone SQL script is also provided in db/promptvault.sql for reference / manual setup.

Reset the database

docker rm -f promptvault-mysql
# then re-run step 1 and step 2