PromptShield 🛡️

Zero-dependency PII & Secret Redactor for LLM Prompts and AI Pipelines. Guard your enterprise data before sending prompts to OpenAI, Anthropic, or Gemini.

---

⚡ Features

• 🔒 Secrets & API Key Redaction: Automatically detects OpenAI keys, GitHub tokens, AWS keys, JWTs, and Bearer tokens.
• 👤 PII Data Protection: Redacts emails, phone numbers, credit card numbers, SSNs, and IPv4 addresses.
• 🚀 Zero Dependencies: Lightweight, ultra-fast regex engine written in strict TypeScript.
• 🛠️ Configurable: Easily add custom domain patterns or toggle specific redaction categories.

---

📦 Installation

�ash npm install promptshield

---

🚀 Quick Start

` ypescript import { PromptShield } from 'promptshield';

const shield = new PromptShield();

const unsafePrompt = User email: john.doe@company.com API Key: sk-proj1234567890abcdef1234567890 AWS Key: AKIAIOSFODNN7EXAMPLE ;

const result = shield.redact(unsafePrompt);

console.log(result.redactedText); /* Output: User email: [REDACTED_EMAIL] API Key: [REDACTED_OPENAI_KEY] AWS Key: [REDACTED_AWS_KEY] */

console.log(result.hasPII); // true console.log(result.hasSecrets); // true console.log(result.redactedCount); // 3 `

---

📄 License

This project is licensed under the MIT License.