Prompt Injection Detection Engine (PIDE)

An enterprise-grade, multi-layered security gateway designed to intercept, analyze, and mitigate prompt-injection attacks against Large Language Models (LLMs).

---

Overview

Prompt Injection Detection Engine (PIDE) is a high-performance security layer built to safeguard LLM applications from malicious prompt manipulation. Moving beyond single-point validation, PIDE executes a robust, multi-tier pipeline incorporating a lightning-fast rule-based Layer 1 regex engine, embedding similarity checks, heuristic analysis, and dynamic risk scoring.

---

Core Architecture & Detection Layers

Layer / Module	Mechanism & Logic	Objective
Layer 1	Rule-Based Regex Engine	Instant pattern matching for known injection signatures ($O(N)$).
Layer 2	Embedding Similarity & FAISS	Vector-space semantic distance analysis against known malicious exemplars.
Layer 3	Heuristic & Risk Scoring	Behavioral weighting and dynamic multi-variable scoring thresholds.
API Gateway	FastAPI REST Gateway	Low-latency asynchronous request handling and secure wrapper integration.
Playground UI	Gradio Interactive Interface	Real-time monitoring, live auditing, and payload testing dashboard.

---

Application Interface & System Preview

1. PIDE Main Interface

2. Interactive Playground & Detection Engine

3. Execution Logs & Detailed Threat Output

4. Live Pipeline Simulation Sample

---

Quick Start (No Extra Files Needed)

1. Clone the Repository

git clone [https://github.com/AyeshaaRafaqat/Prompt-Injection-Detection-Engine.git](https://github.com/AyeshaaRafaqat/Prompt-Injection-Detection-Engine.git)
cd Prompt-Injection-Detection-Engine