A protocol meets the command line.

What is it?

A command-line client that exposes MCP server tools as terminal commands with JSON output.

What problem does it solve?

Call MCP tools from a shell. A small bridge with an unusually useful destination.

What does it replace?

Potentially small, hand-written MCP clients and one-off integration scripts.

What doesn’t it replace?

The server, its permissions or the service behind it. It does not make an unsafe tool safe.

Who is it for?

Developers who like pipes and scripts and already have an MCP server to call.

The review notebook

A tool does not need to be a destination to be useful. Sometimes the best contribution is making an existing capability available where work already happens. JSON output brings MCP calls within reach of ordinary shell processing.

The first thing we would inspect in use is failure. Scripts need to distinguish an empty answer from a rejected call, a timeout or expired authentication. That distinction matters more than how short the example command looks.

Known catches & open questions

Authentication, server compatibility and failure handling remain part of the job. A successful happy-path call is only the beginning.

Current judgement

This is a promising lead, not a recommendation. We will only make a hands-on judgement after the stated test has been run.

Next on the bench

Compare a successful call, an invalid argument, a timeout and denied access; inspect exit codes and output for each.

Sources & evidence

Project repository and README ↗
Release history ↗

Public material checked 5 September 2026. Version labels are the latest stable release returned at that check, where available. Platform descriptions are upstream claims, not Toolglass acceptance tests.