os-control-mcp
Allows creation of persistent terminal sessions using the Zsh shell, with capabilities to send input and read incremental output.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@os-control-mcprun 'npm install' in a new bash session and show the output"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
os-control-mcp
Local MCP for persistent terminal sessions. The process keeps Bash, CMD, PowerShell, pwsh, or WSL active and separates input sending from incremental output reading.
It also has persistent goals: the agent can create a goal with acceptance criteria, run multiple steps, verify evidence, and continue until it completes, fails, gets blocked, or reaches the defined limits.
Warning: this server controls processes with the permissions of the user who runs it. Use it locally, do not run as administrator/root by default, and never expose the process directly to the internet.
Transport
The server uses MCP JSONL over stdio. It does not open a network port by default.
Related MCP server: Persistent Terminal MCP Server
Shells
bash,sh,zshon POSIX systems;cmd,powershell, andpwshon Windows;wsl-bashon Windows, using the distribution defined inOS_CONTROL_WSL_DISTROorkali-linuxby default.
Interactive flow
os_create_sessioncreates a persistent session.os_send_inputwrites immediately to the PTY/ConPTY.os_read_outputusessince_seqto fetch only new output.os_session_statustracks PID, shell, directory, and termination.
Outside of an autonomous goal, potentially destructive commands require
explicit confirmation. In an autonomous goal, authorization is continuous within
the step and time limits defined by the goal. The server is intended for
local use and should remain protected by authentication and scoping when
placed behind a remote tunnel.
Autonomous goals
Recommended flow:
os_create_sessioncreates the persistent terminal.os_goal_startreceives the goal, criteria, andsession_id.os_send_inputuses thegoal_idto run without repetitive confirmation.os_read_outputreads the incremental output.os_goal_updaterecords action, result, evidence, and next step.The agent repeats until all criteria are proven.
The server does not invent the next action or call the model on its own; it maintains the state and limits of the goal. The client/agent must keep making calls until the terminal state.
Installation
git clone https://github.com/Willian-2-0-0-1/os-control-mcp.git
cd os-control-mcp
python -m pip install -r requirements.txtOn Windows, pywinpty enables ConPTY for CMD and PowerShell. On Kali/Linux,
the backend uses the native PTY. The .mcp.json file can be used by MCP
clients that accept local servers over stdio.
Plugin for GPT/Codex
The repository contains the .codex-plugin/plugin.json manifest and the MCP server.
This allows distributing the package to clients compatible with MCP plugins.
To use with ChatGPT with a local server, each user must run their own instance and configure a private connection. A GitHub repository, by itself, does not give ChatGPT access to the user's computer.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Operate Linux, macOS and Windows from your LLM. Every action runs through an auditable allowlist.
Adaptive plan/build/review cycles for AI coding assistants, persisted across sessions.
Hosted runtime for persistent agent teams, durable workflows, memory, schedules, and goals.
Persistent work tracking for AI agents: tasks, status and history that follow you across machines
Related MCP Servers
- AlicenseBqualityDmaintenanceExecute terminal commands locally or remotely via SSH with session persistence and environment variable support. Manage terminal sessions that maintain state for up to 20 minutes, enabling efficient command execution workflows. Connect using stdio or SSE for flexible integration with AI models and a12MIT
- AlicenseNot gradedqualityCmaintenanceEnables AI assistants to create and manage persistent terminal sessions that continue running even after disconnection, with features like smart output management, web UI visualization, and automatic bug fixing with Codex integration.34MIT
- FlicenseNot gradedqualityDmaintenanceProvides stateful, interactive terminal access for LLMs to spawn and maintain persistent processes like SSH sessions, debuggers, and REPLs with continuous input/output interaction across commands.8-
- AlicenseNot gradedqualityFmaintenanceEnables LLMs to create and manage persistent, interactive shell sessions with full terminal emulation and PTY support. It allows for sequential command execution and supports interactive programs like vim or htop through specialized streaming and snapshot output modes.4MIT