Security Testing for AI Agent Skills

Automatically test your AI agent skills for prompt injection, data exfiltration, and other security vulnerabilities. Get actionable reports in minutes.

How It Works

Paste GitHub URL

Submit any public or private GitHub repository containing your AI agent skills.

Select Skills

Choose which skill folders to test. We auto-detect skill directories in your repo.

Get Security Report

Our multi-agent system tests your skills and delivers a detailed severity report.

Features

Multi-Agent Testing

Multiple specialized security agents probe your skills for prompt injection, data exfiltration, and more.

SSRF Protection

Enterprise-grade URL validation with DNS rebinding prevention and IP pinning.

Commit-Based Dedup

Never re-test the same code. Results are cached per commit SHA across the platform.

Severity Reports

Clear severity ratings from CLEAN to CRITICAL with detailed evidence and remediation guidance.

Frequently Asked Questions

Ready to secure your AI agent skills?

Start testing for free. No credit card required.