g0

g0 is an open-source security control layer for AI agents created by Guard0. It provides comprehensive security assessment capabilities across the entire AI agent lifecycle — from discovery and assessment to testing, monitoring, and compliance. With 1,200+ security rules mapped across 12 domains and 10 supported frameworks, g0 enables teams to answer three critical questions: what agents do you have, what can they access, and can you prove they're under control? The tool includes 4,000+ adversarial payloads for testing, generates AI Bills of Materials (AI-BOM) in CycloneDX format, and provides fleet-wide monitoring through a background daemon. g0 integrates directly into CI/CD pipelines with SARIF output support for GitHub Code Scanning, enabling automated quality gates that prevent insecure agents from shipping.

Use this tool if you:
• Need to discover and inventory all AI agents in your environment
• Want to assess agent security posture against industry frameworks
• Need to test agents against adversarial attacks and prompt injection
• Must generate AI Bills of Materials for compliance and supply chain security
• Want to integrate agent security checks into your CI/CD pipeline
• Need continuous monitoring for agent fleet security and drift detection
• Must comply with EU AI Act, NIST, ISO, or OWASP standards