claude-codex-settings
My personal Claude Code and OpenAI Codex setup with battle-tested skills, commands, hooks, agents and MCP servers that I use daily.
claude mcp add --transport stdio fcakyon-claude-codex-settings node server.js \ --env PORT="Optional port if needed by the MCP server" \ --env CLAUDE_API_KEY="Placeholder for Claude API key if required"
How to use
This MCP server provides the Claude Codex Settings integration for managing Claude Code plugins and related configuration within the MCP ecosystem. It leverages the Claude Codex workflow to coordinate settings, synchronization of CLAUDE.md content, and allowlists across the CLAUDE Codex setup. With this MCP server active, you can leverage the associated MCP endpoints to organize and apply configuration changes to your Claude Code environment, and coordinate between CLAUDE.md instructions, the permissions allowlist, and context-refresh workflows. The repository’s documentation indicates a broad plugin system (Azure MCP, GCloud MCP, MongoDB MCP, Slack MCP, etc.) and this MCP server focuses on centralizing Codex-specific settings, synchronization tasks, and hooks that support consistent behavior across plugins and skills. Use the server to access or trigger settings synchronization, context refresh routines, and any Codex-specific configuration commands exposed by the MCP layer.
How to install
Prerequisites:
- Node.js installed on the host where you run the MCP server (recommended LTS version).
- Access to the Claude Code environment or a compatible MCP runner that supports the MCP protocol.
- Basic familiarity with running Node.js applications and starting an MCP server.
Installation steps:
-
Prepare the environment
- Ensure Node.js is installed: node -v and npm -v should return version numbers.
- Create a project directory for the MCP server or clone the repository that contains claude-codex-settings.
-
Install dependencies (if a package.json exists)
npm install
- Run the MCP server
node server.js
- Verify the server is up
- Check the console for a listening port (e.g., Listening on port 3000) or use curl to hit the MCP endpoint.
- If a specific port is required by your MCP environment, set PORT accordingly:
PORT=3000 node server.js
- Integrate with your MCP orchestrator
- Register the server under the MCP framework using your platform’s configuration (e.g., add claude-codex-settings as a server in your MCP registry).
- If applicable, ensure any required environment variables (like CLAUDE_API_KEY) are provided in your deployment environment or via your orchestration tooling.
Additional notes
Tips and common considerations:
- The CLAUDE_API_KEY (or equivalent) may be required if the server calls out to Claude services for synchronization or content operations. Do not commit secrets; use your environment’s secret management.
- If you modify environment variables, restart the MCP server for changes to take effect.
- Review any relevant llms.txt or plugin documentation referenced in the repository README to understand how CLAUDE.md synchronization interacts with MCP flows.
- Ensure compatibility with your Claude Code plugins set, particularly those related to CLAUDE.md synchronization and allowlists.
- If you encounter port conflicts, adjust the PORT variable or configure the MCP orchestrator to route to a dedicated port for this server.
Related MCP Servers
ios-simulator-skill
An IOS Simulator Skill for ClaudeCode. Use it to optimise Claude's ability to build, run and interact with your apps, without using up any of the available token/context budget.
codexia
Agent Workstation for Codex CLI + Claude Code — with task scheduler, git worktree & remote control
claude-emporium
🏛 [UNDER CONSTRUCTION] A (roman) claude plugin marketplace
kindly-web-search
Kindly Web Search MCP Server: Web search + robust content retrieval for AI coding tools (Claude Code, Codex, Cursor, GitHub Copilot, Gemini, etc.) and AI agents (Claude Desktop, OpenClaw, etc.). Supports Serper, Tavily, and SearXNG.
mcp-image
MCP server for AI image generation and editing with automatic prompt optimization and quality presets (fast/balanced/quality). Powered by Gemini (Nano Banana 2 & Pro).
sugar
🍰 Sugar - The autonomous layer for AI coding agents