iterm-mcp
MCP server for iTerm: the agent runs commands in the active terminal session, reads output and sends control characters
- 571 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Runs arbitrary commands in your terminal
- No built-in command safety checks
- Can start and interrupt any process
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/iterm-mcpDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Runs arbitrary commands in your terminal
- No built-in command safety checks
- Can start and interrupt any process
This is third-party code. Review the repository files before installing.
What it does
The server gives an agent access to your active iTerm session on macOS. The agent writes commands to the terminal, reads a chosen number of recent output lines and sends control characters like ctrl-c and ctrl-z. This lets it work with REPLs and follow long-running commands while showing the model only the part of the output that matters. You share the terminal with the model and see every step. It installs via npx and targets Claude Desktop and other MCP clients, with few dependencies.
Who it is for. For developers and devops engineers on macOS who like delegating terminal steps to an agent.
Good fit when
- You want the agent to run commands in your iTerm session
- You need to work with REPLs and interactive processes
- You need to interrupt commands with control characters
Not a fit when
- You are not on macOS or do not use iTerm2
- You cannot allow an agent to run commands directly without limits
- You need built-in command safety checks; there are none here
Example request
Run the tests in the current terminal and show the last lines of outputMCP
- Transport
- stdio
- Authentication
- not required
Limitations
You need iTerm2 and Node.js version 18 or newer. The server does not evaluate command safety: it runs whatever the agent decides, and responsibility is on the user. For multi-step tasks you may need to interrupt the model if it goes off track. It is best to start with small tasks.
How to disable. Remove the iterm-mcp entry from your MCP client config.
FAQ
Does the server check commands for danger?
No. There are no built-in restrictions; safety is on the user. Watch the activity and interrupt when needed.
Does it work on Windows or Linux?
No, the server is tied to iTerm2 and runs on macOS.
Discussion
No questions or reviews yet.
Related
MCP serversA CLI and TypeScript runtime for calling MCP servers from a terminal, scripts and generated typed clients
A macOS terminal with the termio CLI and a skill: agents spawn sibling sessions and you see which one needs your answer
term-cli runs interactive programs for an agent inside tmux sessions, while a human enters passwords and MFA through term-assist.
An MCP server that lets an agent read files, search code and run commands in a workspace
Hosted MCP server that links claude.ai, ChatGPT and Cursor to your computer's files and terminal over OAuth
A command-line debugger for MCP servers from Giant Swarm: a REPL for tools, OAuth 2.1 and an MCP server mode for Cursor
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| iterm-mcp (this card) | no | not required | 571 | |
| kubectl-mcp-server | no | not required | 972 | reads cluster state and diagnoses pod failures |
| Unreal Engine MCP Server | no | not required | 1,033 | Unreal Engine assets, actors and blueprints |
| Android MCP | no | not required | 924 | controls an Android device through ADB |
README in short
The README presents a server that gives the model access to an iTerm session with three tools: write to the terminal, read a set number of output lines, and send a control character. It flags safety: there are no built-in restrictions and the user is responsible for execution. Install is via npx, with a Smithery option, and it needs iTerm2 and Node.js 18 or newer. It lists build, development and MCP Inspector debugging commands. MIT licensed.