provider-scanner

What problem does it solve?

This Skill safeguards AI communication by detecting and preventing malicious inputs, sensitive data leaks, and unauthorized access through advanced scanning mechanisms.

Core Features & Use Cases

• Input Security: Scans inbound messages for prompt injection, jailbreaks, and other adversarial attacks.
• Output Security: Monitors outbound agent responses for Personally Identifiable Information (PII), credentials, and sensitive data exfiltration.
• Canary Tokens: Generates and checks for unique tokens to detect data leakage in agent outputs.
• Use Case: Ensure that user queries cannot manipulate the AI into revealing confidential system information or executing harmful commands, and that the AI's responses do not inadvertently expose user data.

Quick Start

Use the provider-scanner skill to add a new regex pattern to the input scanners.