Scan Your AI Agent — Free, Instant, Anonymous

How The Scan Works

Our AI Agent Security Scanner performs 27 comprehensive checks across 6 critical security categories:

• Permissions & Access Control — exec security modes, sandbox settings, tool restrictions
• Prompt Injection Defense — system prompt protections, input validation
• Data Exfiltration Prevention — allowlists, SSRF policies, network restrictions
• Supply Chain Integrity — plugin allowlists, dependency verification
• Sub-Agent Safety — sandboxing, trust scoring, behavioral inheritance
• Audit & Logging — logging configuration, monitoring, telemetry

Each check is weighted by severity to compute your overall security score. Scores of 80+ qualify for Pitstop Certification.

Want a Deeper Scan?

Run the full CLI version for advanced checks, custom policies, and integration with your CI/CD pipeline:

📖 Read the full documentation at /beta