IBM MQ MCP Server

This project aims to give AI clients a safe, observable interface to IBM MQ without assuming that queue managers run on Kubernetes or are managed by a particular operator.

The design is intentionally capability-oriented: each named connection profile grants only the operations its operator permits. A production profile may be read-only while a development profile permits message production or selected administrative changes.

Start with the quickstart, feature scope, and proposed system.

Operator docs cover deployment, observability, security/threat model, and the provisional IBM MQ version matrix.