Discover and explore top open-source AI tools and projects—updated daily.
Yeachan-HeoMulti-agent orchestration for OpenAI Codex CLI
New!
Top 41.1% on SourcePulse
OmX (Oh My Codex) provides a multi-agent orchestration layer for the OpenAI Codex CLI, transforming it from a single-session tool into a coordinated system. It targets developers and power users working with Codex CLI on larger projects, offering structured decomposition, staged execution, and persistent memory to enhance productivity and manage complex workflows.
How It Works
OmX acts as an add-on, leveraging Codex CLI's extension points. It introduces role-based agents (via /prompts), repeatable workflow skills (via $name), and team orchestration within tmux. The system manages persistent state and memory through MCP servers, enabling structured work through decomposition and staged execution (e.g., team-plan -> team-exec -> team-verify). This approach provides a lifecycle for modes, memory surfaces, and operational controls for launch, verification, and cancellation, adding structure for complex tasks beyond single-session interactions.
Quick Start & Requirements
npm install -g oh-my-codexnpm install -g @openai/codex), Codex authentication configured.omx setup to install prompts, skills, and configuration. omx doctor can diagnose installation and runtime issues.Highlighted Details
AGENTS.md into Codex launches.--madmax for bypassing safety checks in trusted environments.Maintenance & Community
The project is inspired by oh-my-claudecode and adapted for Codex CLI. Specific details on active contributors, sponsorships, or community channels (like Discord/Slack) are not detailed in the provided README.
Licensing & Compatibility
The project is released under the MIT license, which is permissive for commercial use and integration into closed-source projects.
Limitations & Caveats
The tool requires macOS or Linux (or WSL2 on Windows). It is dependent on the Codex CLI and its authentication setup. The --madmax launch flag bypasses Codex sandbox approvals and should only be used in trusted environments. In team mode, users are cautioned against shutting down active tasks unless aborting.
22 hours ago
Inactive