Guide · Agent skill

How to use the MCPorter CLI Helper agent skill

Updated 2026-10-29 · written with AI from the project's own documentation and checked against our catalogue data

The mcporter agent skill teaches an AI agent to use the mcporter command-line tool, which talks to MCP (Model Context Protocol) servers. MCP servers expose tools such as issue trackers or web scrapers, and this skill lets your agent list those tools, check their inputs, and call them over HTTP or local stdio. This guide is for developers who already use MCP servers and want an agent to work with them, and it explains how to install the skill, trigger it, and use it safely.

MCPorter CLI Helperspinabot/brigade · MIT Download the skill

Quick start

1Install Node.js and mcporter. Install Node.js, then install the mcporter package, which provides the mcporter CLI.
2Download and unzip the skill. Download the skill ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder.
3Or upload the ZIP in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
4Confirm your servers are reachable. Run mcporter list to see the servers and tools you can access.
5Inspect a tool's inputs. Run mcporter list <server> --schema to see what arguments a tool expects.
6Ask the agent for a task. Describe the server, tool and values in plain language, and review any write action before it runs.

What is an agent skill and what does mcporter add?

An agent skill is a folder with a SKILL.md file that gives an AI agent instructions for a specific task. Without a skill, the agent does not know a tool's commands or syntax. This skill is the instruction set for the mcporter CLI.

With it, the agent can list MCP servers and their tools, read tool schemas, call tools with key=value arguments or a JSON payload, handle OAuth login, edit a config file, and generate CLI or TypeScript wrappers around MCP servers.

Install the mcporter skill

Download the skill ZIP from this page and unzip it. For a personal install, place the folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.

In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.

The skill needs the mcporter CLI itself. The source describes it as a Node package named mcporter, so Node.js is required for that install step.

Personal: ~/.claude/skills/
Project: .claude/skills/ inside your project
Claude apps: Settings → Capabilities → Skills (upload the ZIP)

Trigger the skill with natural requests

You do not need special syntax. Ask the agent in plain language to work with an MCP server, for example to list the tools a server offers, check a tool's inputs, or call a tool with specific values. The skill covers the mcporter commands behind those requests.

Name the server and the tool when you can. If the agent needs more detail, it can run mcporter list to discover what is available.

List what a server offers: list the tools on your linear server
Check inputs: show the schema for a tool before calling it
Call a tool: create or fetch something with specific arguments

Core mcporter commands the agent uses

The main commands are listing, schema inspection and calling. Calls accept a selector form, a function-style form, a full URL, a stdio server, or a JSON payload.

List servers: mcporter list
Show a server's schema: mcporter list <server> --schema
Call with a selector: mcporter call linear.list_issues team=ENG limit:5
Call a stdio server: mcporter call --stdio "bun run ./server.ts" scrape url=https://example.com
Call with JSON arguments: mcporter call <server.tool> --args '{"limit":5}'

mcporter list
mcporter list <server> --schema
mcporter call <server.tool> key=value

Handle auth, config and the daemon

Servers that need OAuth are handled with mcporter auth. The default config file is ./config/mcporter.json, and you can point to another file with --config. The config subcommands are list, get, add, remove, import, login and logout. The source also documents a daemon with start, status, stop and restart.

Servers that need OAuth logins or API credentials still need those credentials from you. The skill does not create them.

OAuth: mcporter auth <server | url> [--reset]
Config: mcporter config list|get|add|remove|import|login|logout
Daemon: mcporter daemon start|status|stop|restart

mcporter auth <server | url> [--reset]

Generate CLI or TypeScript wrappers

The skill can generate a command-line wrapper for an MCP server, inspect an existing generated CLI, or emit TypeScript for a server, either as a client or as types. Generated CLIs can be built from a configured server name or from a server URL.

Use --output json when a script or the agent needs machine-readable results.

Generate CLI from a configured server: mcporter generate-cli --server <name>
Generate CLI from a URL: mcporter generate-cli --command <url>
Inspect a generated CLI: mcporter inspect-cli <path> [--json]
Emit TypeScript: mcporter emit-ts <server> --mode client|types

mcporter generate-cli --server <name>
mcporter inspect-cli <path> [--json]
mcporter emit-ts <server> --mode client|types

Safety notes before you let it run

This skill runs scripts and uses the network. Calls go to real MCP servers, so the agent can change data on them. Review any call that creates, updates or deletes something before it runs, especially on production servers.

Only use the skill with servers you trust and with credentials you are comfortable letting an agent use.

Review write actions, such as creating an issue, before the agent calls them
Test against a non-production server first
Check which config file is active, since the default is ./config/mcporter.json

What people use it for

List the tools an MCP issue tracker exposes before building an integration
Call a web scraper MCP tool on a URL during development
Test an MCP tool from the terminal before wiring it into an app
Generate a CLI or TypeScript wrapper around an MCP server
Edit or import entries in the mcporter config file

Frequently asked questions

What is the mcporter agent skill?

It is an agent skill that teaches an AI agent to use the mcporter command-line tool to list, configure, authenticate and call MCP server tools. It documents the mcporter commands so the agent knows the syntax.

How do I install the mcporter skill?

Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use or a project's .claude/skills/ folder. You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.

What do I need to run it?

You need the mcporter CLI, which the source describes as an npm package named mcporter, so Node.js is needed for that install. You also need access to the MCP servers you call, including any OAuth logins or API credentials they require.

Is the mcporter skill free?

The skill is listed under the MIT licence. This page does not state pricing for the skill or the Claude apps, so check those separately.

Does the skill set up MCP servers for me?

No. The skill only documents the mcporter CLI. It does not install or set up the MCP servers themselves.

Is it safe to let the agent call tools?

The skill runs scripts and uses the network, and the agent can make changes on the server. Review write actions such as creating issues before they run, and test on non-production servers first.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use this skill if you already run or use MCP servers and want an agent to list, inspect and call their tools, or to generate wrappers from them. Skip it if you do not use MCP servers or the mcporter CLI, since the skill does not set those up.

Download the skill