kvlar-proxy

MCP security proxy — intercepts Model Context Protocol messages and evaluates them against loaded security policies before forwarding.

This crate provides the runtime enforcement layer. It sits between an AI agent and its tool servers, ensuring every tool call passes through the Kvlar policy engine before execution.

Architecture

Agent ──► kvlar-proxy ──► MCP Tool Server
              │
              ├── kvlar-core (policy evaluation)
              └── kvlar-audit (structured logging)