oh-my-openagent plugin
Orchestrator plugin for OpenCode and Codex CLI: multiple agents, hooks, built-in MCPs, ultrawork mode and hash-anchored edits
- 70,440 ★ on the repository
- Last commit
Broad access
Why this level
- Autonomous modes run commands and change code without per-step approval
- The light edition can enable broader Codex permissions
Install
- Works in
Manual install
bunx oh-my-openagent installThe wizard registers the plugin in opencode.json and sets up models. Add --platform=both to install for both hosts.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Autonomous modes run commands and change code without per-step approval
- The light edition can enable broader Codex permissions
This is third-party code. Review the repository files before installing.
What it does
oh-my-openagent (OmO) extends OpenCode and Codex CLI with agents, lifecycle hooks, slash commands and skills. Typing ulw or mass ulw in a prompt switches on a mode where the agent plans and drives the task to completion, spreading work across models. Skills bring their own MCP servers that start only for the task, keeping the context clean. The optional Hashline mode tags each line read with a hash, and an edit is rejected if the file changed since. Built-in skills cover git, browser work, frontend, refactoring and security audits.
Who it is for. For OpenCode or Codex CLI users who want a more autonomous multi-model workflow.
Good fit when
- You want the agent to plan and finish a large task on its own
- You want several models in one session
- Agent edits often fail because file contents went stale
Not a fit when
- You use Claude Code or Cursor, which are not supported
- You need predictable step-by-step control over every edit
Example request
ulw migrate the payments module to the new SDK and run all testsLimitations
The full edition works only with OpenCode, the light edition with Codex CLI, and the standalone Senpi build is beta. The license is custom rather than a standard open source one; read LICENSE.md. Autonomous modes consume many tokens from your providers.
How to disable. Remove the plugin from opencode.json or its blocks from ~/.codex/config.toml. For the Senpi edition run npm uninstall -g omo-ai.
FAQ
Which edition should I pick?
Ultimate for OpenCode, Light for Codex CLI, and the Senpi beta via npm i -g omo-ai@beta if you want no host.
Can I install the omo package from npm?
No, the README warns that omo is an unrelated package by another author.
Discussion
No questions or reviews yet.
Related
PluginsAn OpenCode plugin: an orchestrator hands work to six specialists, each on its own model, and merges their results into one answer
A multi-agent harness that verifies results by artifacts: stop hooks, an independent judge and an event log for Claude Code, Codex, Cursor and more
An open source coding agent for the terminal and desktop with build and plan modes
oah links 80-plus skills, five roles and /commit, /review, /test-tdd commands into Claude Code, Cursor, Codex and other agent folders
A self-hosted control center for coding agents: OpenHands, Claude Code, Codex and Gemini on local and remote backends
A local context store for coding agents that captures project knowledge and reuses it across Codex, Claude Code and OpenCode
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| oh-my-openagent (this entry) | no | 70,440 | |
| oh-my-opencode-slim | no | 9,844 | an OpenCode plugin |
| OpenCode Swarm | no | 514 | OpenCode plugin that turns a session into a team of agents led by an |
| FableCodex | no | 439 | a Codex plugin with an evidence-backed goal ledger and a findings gate |
README in short
The README describes three editions: Ultimate for OpenCode with eleven agents, hooks and four built-in MCPs, Light for Codex CLI with portable components, and Senpi as a standalone beta. A table shows the install command for each and what lands on disk. Sections explain skill-embedded MCPs, Hashline edits and deep project initialization.