agent-firewallSecurityAutonomous AgentsShubhbhangooAlicense-Not gradedqualityAmaintenanceA policy-based security layer for AI agents and MCP tools that evaluates tool requests against security policies before allowing, denying, or requiring human approval for execution. Updated 22 days ago (2026-09-14 20:08 UTC)1MIT