AI Security Scorecard

The AI Security Scorecard is a public database of security audits for AI tools and open source projects.

Each tool is scored across five metrics on a scale of 0-100:

DEPENDENCY SECURITY
Analysis of third-party dependencies, known CVEs, and outdated packages.

NETWORK SAFETY
Examination of API calls, external endpoints, and potential data exfiltration.

PERMISSION MODEL
Review of filesystem access, system permissions, and privilege escalation risks.

PRIVACY COMPLIANCE
Assessment of data collection, telemetry, and GDPR implications.

LICENSE COMPLIANCE
Verification of license compatibility and legal usage restrictions.

OVERALL SCORE
Weighted average of all five metrics. Tools scoring below 60 are considered high risk.

Security reports are published regularly. Subscribe to receive new audits directly in your inbox.