🛡️ AuditFlow — Enterprise AI Workflow Versioning & Audit Platform

AuditFlow is an enterprise-grade platform for versioning, auditing, and validating AI agent workflows. It provides immutable version control for LLM prompts and completions, automated 4-layer security validation gates, semantic diffing between workflow editions, Role-Based Access Control (RBAC), and automated AWS S3 export backup capabilities.

---

🌟 Key Features

📚 1. Immutable Workflow Versioning

• Chapters: Individual execution steps tagged with actor, prompt, model (gpt-4o, claude-3.5-sonnet), temperature, seed, and SHA-256 validation status.
• Books: Version-controlled sequences of chapters representing full AI agent execution pipelines.
• Editions: Linear parent-child versioning (v1 → v2) ensuring complete historical auditability.
• Feature Shelves: Logical categorization of workflows by feature (e.g. finance, support, security).

🔍 2. Visual Semantic Diff Engine

• Compare prompt and completion changes side-by-side between any two workflow editions (Book v1 vs Book v2).
• Identifies Kept, Added, and Removed execution steps with color-coded diff syntax highlighting.

🛡️ 3. 4-Layer Validation Gate Harness

• Layer 1 (Schema & Format): Validates JSON output structure and required keyword presence.
• Layer 2 (ReDoS Regex): Evaluates regex compliance with ReDoS attack protection.
• Layer 3 (PII Privacy): Scans input prompts and completions for exposed sensitive PII data.
• Layer 4 (Security Scanner): Detects prompt injection patterns (bypass, ignore instructions, system prompt leak attempts).

🔐 4. Stateless JWT Authentication & 4-Tier RBAC

• Stateless HMAC-SHA256 JWT authorization with 24-hour token expiration.
• Interactive authentication modal with Sign In, Sign Up, and One-Click Demo Account Switching:
  • 🔴 ADMIN: Full system control, user management, API key revocation, and system backups.
  • 🟣 DEVELOPER: Chapter creation, workflow assembly, and validation testing.
  • 🟢 AUDITOR: Compliance inspection, semantic diff viewing, and audit logs.
  • ⚪ VIEWER: Read-only observation across feature shelves.

☁️ 5. Cloud Backups & AWS S3 Sync

• Online point-in-time database snapshot packaging.
• On-demand and scheduled export metadata syncing to AWS S3 buckets (s3://auditflow-exports-prod/backups/).

---

🏗️ Architecture & Technology Stack

               ┌─────────────────────────────────────────┐
               │    React 19 + TypeScript + Tailwind v4   │
               │            AuditFlow Dashboard          │
               └────────────────────┬────────────────────┘
                                    │ HTTP / REST API
                                    ▼
               ┌─────────────────────────────────────────┐
               │       Spring Boot 3.4.3 (Java 21)       │
               │   Spring Security + JJWT Auth Filter    │
               └────────┬───────────────────────┬────────┘
                        │                       │
                        ▼                       ▼
               ┌─────────────────┐     ┌─────────────────┐
               │ PostgreSQL / H2 │     │   Redis 7.0     │
               │  Relational DB  │     │ Cache & Sessions│
               └─────────────────┘     └─────────────────┘

Component	Technology
Backend	Java 21, Spring Boot 3.4.3, Spring Security 6, JJWT 0.12.6, PostgreSQL, Redis, Lombok
Frontend	React 19, TypeScript 5.7, Vite 6, Tailwind CSS v4 (@tailwindcss/vite), Lucide React
Testing	JUnit 5, Mockito, Spring Security Test (14/14 automated tests passed)
Containerization	Docker, Multi-Stage Dockerfile, Docker Compose, Nginx

---

⚡ Quick Start

Option A: Launch with Docker Compose (Recommended)

1. Clone the repository:

   git clone https://github.com/ATHARVA262005/AuditFlow.git
   cd AuditFlow

2. Spin up the full production stack (Backend + Frontend + Postgres + Redis):

   docker-compose up --build

3. Access the dashboard:

   • Frontend UI: http://localhost:80
   • Backend API: http://localhost:8080/api/v1

---

Option B: Local Development Setup

1. Backend (Java 21 + Spring Boot)

cd backend
mvn clean package -DskipTests
mvn spring-boot:run

Backend runs on http://localhost:8080.

2. Frontend (React 19 + Vite)

cd frontend
npm install
npm run dev

Frontend runs on http://localhost:5173.

---

🔑 Pre-Seeded Demo Credentials

You can test real-time Role-Based Access Control using the built-in preset accounts:

Role	Username	Password	Email	Permissions
ADMIN	admin	admin123	admin@auditflow.io	Full Access, RBAC & Key Management
DEVELOPER	developer	dev123	developer@auditflow.io	API Key Generation, Validation Testing
AUDITOR	auditor	audit123	auditor@auditflow.io	Diff Viewing & System Audit Logs
VIEWER	viewer	viewer123	viewer@auditflow.io	Read-only observation

---

🧪 Automated Testing

Run the full backend unit test suite:

cd backend
mvn test

Test Results

[INFO]  T E S T S
[INFO] Running com.auditflow.service.AuthServiceTest — 4 tests passed
[INFO] Running com.auditflow.service.BookServiceTest — 2 tests passed
[INFO] Running com.auditflow.service.ChapterServiceTest — 2 tests passed
[INFO] Running com.auditflow.service.DiffServiceTest — 1 test passed
[INFO] Running com.auditflow.service.SystemExportServiceTest — 2 tests passed
[INFO] Running com.auditflow.validation.ValidationEngineTest — 3 tests passed
[INFO] Results: Tests run: 14, Failures: 0, Errors: 0, Skipped: 0
[INFO] BUILD SUCCESS

---

📖 API Endpoint Reference

Method	Endpoint	Description	Access Level
POST	/api/v1/auth/register	Register new user account	Public
POST	/api/v1/auth/login	Authenticate user & issue signed JWT	Public
GET	/api/v1/chapters	List all chapters	Authenticated
POST	/api/v1/chapters	Log new chapter step	DEVELOPER / ADMIN
GET	/api/v1/books	List workflow books	Authenticated
POST	/api/v1/books	Create new workflow edition	DEVELOPER / ADMIN
GET	/api/v1/diff/books	Calculate semantic diff between books	AUDITOR / ADMIN
POST	/api/v1/validation/test	Test prompt/completion on 4-layer engine	DEVELOPER / ADMIN
POST	/api/v1/system/s3-sync	Export database telemetry & sync to AWS S3	ADMIN

---

🚀 AWS EC2 & S3 Deployment

See the complete deployment guide in aws-deployment.md for deploying AuditFlow to AWS EC2 with automatic export backups to S3.

---

📄 License & Proprietary Rights

Copyright © 2026 AuditFlow. All Rights Reserved.

This software and associated documentation files are proprietary and confidential. Unauthorized copying, distribution, modification, or transfer of this software, via any medium, is strictly prohibited.