Table of Contents

What is isolated-pymcp?

A secure, isolated container environment that combines Model Context Protocol (MCP) and Language Server Protocol (LSP) for Python development. It enables LLMs to access powerful code intelligence features while maintaining strict security boundaries, and is designed for developers exploring algorithm analysis, Claude Code integration, and MCP/LSP educational materials.

How to use isolated-pymcp?

Build the Docker/Podman image with make build, then run the container with make run. Verify MCP server connectivity with make test. Analyze algorithms using make analyze ALGO=<algorithm> or make claude-analyze ALGO=<algorithm>. The project also provides custom Claude commands such as /fix-github-issue, /create-pr, and /analyze-algorithm.

Key features of isolated-pymcp

• Container isolation from host system with least access
• MCP and LSP integration for code intelligence
• Advanced algorithm implementations with complexity analysis
• Custom Claude commands for GitHub issue resolution
• Literate programming with org-mode and Makefile automation
• Multiple client interfaces with the same security model

Use cases of isolated-pymcp

• Analyze Python algorithm implementations with complexity analysis and security
• Execute Python code securely in an isolated container via MCP
• Use Claude Code to review and fix GitHub issues from the repository
• Learn MCP and LSP integration through provided educational course materials

FAQ from isolated-pymcp

What is the difference between isolated-pymcp and a standard Python environment?

It provides container isolation, non-root user execution, restricted port exposure, and integrated MCP/LSP capabilities for AI-assisted development, whereas a standard Python environment lacks these security boundaries and protocol bridges.

What are the runtime requirements?

Docker or Podman, Python, Make, and GitHub CLI (gh) for secrets management. The container runs on Alpine Linux with Python 3 support.

Where does user data live?

User code and algorithms reside inside the container. Secrets (e.g., ANTHROPIC_API_KEY, GH_PAT) are managed via GitHub Secrets and passed to the container at runtime.

How does the MCP transport work?

It uses JSON-RPC over stdio. Example: tools/call method with run_python_code argument to execute Python code and capture output.

How does isolated-pymcp handle authentication?

Authentication is handled via GitHub Secrets. Secrets are set up using ./scripts/setup_secrets.sh or manually with gh secret edit, and then injected into the container at build/run time.