Prompt Shield Scan
Version updated for https://github.com/markmishaev76/Prompt-Shield to version v1.0.0.
- This action is used across all versions by ? repositories.
Go to the GitHub Marketplace to find the latest changes.
Action Summary
Prompt Shield is a GitHub Action designed to detect and prevent indirect prompt injection attacks in AI-assisted workflows by scanning issues, pull requests, and comments for malicious content. It automates AI security by leveraging a four-layer defense system to filter, sanitize, detect, and tag untrusted content, addressing a critical vulnerability in AI systems. Key capabilities include high-accuracy detection (97%+), minimal false positives, and customizable threat handling, ensuring secure and reliable AI interactions.
Release notes
🛡️ Prompt Shield
Multi-layered defense against indirect prompt injection attacks.
Quick Start
- uses: markmishaev76/Prompt-Shield@v1.0.0
Features
- 97.62% detection accuracy
- 4-layer defense architecture
- 38+ detection patterns
Full Changelog: https://github.com/markmishaev76/Prompt-Shield/compare/v1.0.1...v1.0.0