AgentGuard
AgentGuard sits between your users and AI agents, providing three critical capabilities: real-time security scanning to block prompt injection attacks
Core trick
Intercepts requests to agents and scans each prompt for injection patterns before forwarding it to the model.