prompt-injection-guard

What problem does it solve?

This Skill protects AI applications from malicious prompt injection attacks, ensuring the integrity and security of AI-driven systems.

Core Features & Use Cases

• Input Validation: Detects and sanitizes potentially harmful user inputs before they reach the LLM.
• Output Validation: Verifies that the LLM's output does not contain sensitive information or unintended instructions.
• Defense Strategies: Implements layered security including blocklists, prompt structuring, canary tokens, and LLM-based detection.
• Use Case: When building a customer-facing chatbot that processes user queries, use this skill to prevent users from manipulating the chatbot into revealing sensitive data or performing unauthorized actions.

Quick Start

Use the prompt-injection-guard skill to validate and sanitize user input before processing it with an LLM.