Development Safety System - MCP Server

What is Development Safety System - MCP Server?

The Development Safety System MCP Server is an MCP (Model Control Protocol) server that provides session continuity and sandbox safety for LLM development workflows. It targets developers using AI-assisted coding tools, enabling them to save and restore complete session context and experiment in isolated sandboxes without risking the main project.

How to use Development Safety System - MCP Server?

Install via pip (clone repository, install dependencies with pip install -r requirements.txt, then pip install -e .). Use an MCP compatible client (e.g., Claude Desktop) to invoke tools such as create_sandbox, save_session_state, load_session_state, sync_to_main, check_activity, commit_progress, and check_mcp_status.

Key features of Development Safety System - MCP Server

• Session state persistence for seamless work continuation
• Sandbox management for isolated development environments
• File-based activity detection to suggest saving state
• Safe synchronization of approved changes to main project
• Automatic backups before any sync operations
• Built-in server health monitoring with heartbeat files

Use cases of Development Safety System - MCP Server

• Create an isolated copy of a project for safe experimentation
• Save complete session context when taking a break
• Resume work exactly where you left off using saved state
• Monitor file activity to prompt saving during inactivity
• Copy approved changes from sandbox back to main project

FAQ from Development Safety System - MCP Server

What is the purpose of this server?

It provides "pick up exactly where you left off" and "never break the main app" functionality for AI-assisted development by managing session state, sandboxes, and safe synchronization.

How do I install it?

Clone the repository, install dependencies with pip install -r requirements.txt, then install the package with pip install -e ..

What are the system requirements?

Python 3.8 or higher, Git (for sandbox version control), and an MCP compatible client such as Claude Desktop.

How can I check if the server is running?

Use the built-in check_mcp_status tool or the external status checker script (python tools/check-mcp-status.py). The server writes a heartbeat file to ~/.dev-safety/mcp_status.json every 30 seconds.

What license is this project under?

MIT License – see the LICENSE file for details.