🛡️ PromptShield — AI & LLM Security Gateway

PromptShield is a lightweight, high-performance security proxy designed to protect Large Language Models (LLMs) from Prompt Injections, Jailbreak exploits, PII data leakage, and malicious intent requests before they hit the upstream model.

---

✨ Key Features

• Prompt Injection & Jailbreak Defense: Detects system prompt overrides, developer-mode triggers, and DAN-style exploits.
• Malicious & Cyber Threat Detection: Blocks unauthorized cyber exploitation commands, hacking tools, and destructive intent payloads.
• Automated PII Masking: Regex-based real-time redaction for Credit Cards, API Keys, and Email addresses ([REDACTED_*]).
• SaaS API Authentication & Rate Limiting: Enforces X-API-KEY verification and per-minute sliding window request throttling.
• Visual Security Dashboard: Interactive dark-mode web console with real-time audit logs and live decision streams.

---

🏗️ Architecture Flow

[ Client / User ] 
       │
       ▼ (Prompt Request)
[ PromptShield Gateway ] ── (Blocked) ──► [ Security Violation Alert ]
       │
       ▼ (Sanitized & Approved)
[ Upstream LLM / AI Agent ] ──► [ Safe Response ]