A local-first CLI and MCP server wrapper that routes shell commands through a unified interface, storing command history locally while compressing terminal output to reduce context noise for AI coding agents.

What it does

PsYcGoD/sage is a Model Context Protocol (MCP) server that acts as a wrapper for shell command execution. It provides a unified interface for AI coding agents to run commands safely and efficiently while maintaining a local history of all executed commands.

How to set up

Install via GitHub repository by cloning PsYcGoD/sage and configuring it as an MCP server. Connect it to your AI coding agent (Claude, ChatGPT, or compatible MCP client) to enable shell command routing through the sage run command interface.

Key features

Local-first architecture with command history storage
Routes all shell commands through sage run for safety and tracking
Compresses terminal output to reduce token usage and context noise
Compatible with multiple AI platforms via MCP protocol
CLI-based configuration and execution
Cross-platform support (Windows, macOS, Linux)

What it does

How to set up

Key features

Local-first architecture with command history storage
Routes all shell commands through sage run for safety and tracking
Compresses terminal output to reduce token usage and context noise
Compatible with multiple AI platforms via MCP protocol
CLI-based configuration and execution
Cross-platform support (Windows, macOS, Linux)

PsYcGoD/sage

What it does

How to set up

Key features

PsYcGoD/sage

What it does

How to set up

Key features

Related Skills

Monday.com MCP Server

Sentry MCP Server

Cloudflare MCP Server