MCP has no security model. I built Heddle — a policy-and-trust layer that turns YAML configs into validated, policy-enforced MCP tool servers. Then I invited AI security audits to break it.

When building Model Context Protocol (MCP) servers or AI agents that consume them, traditional API...

MCP has no security model. I built Heddle — a policy-and-trust layer that turns YAML configs into validated, policy-enforced MCP tool servers. Then I invited AI security audits to…