Oh My OpenCode (now OmO): what it is and how to install it
Oh My OpenCode is now OmO. What changed, how its README installs it, what the install script does, how it carries OpenCode skills over, and its license.
- For
- OpenCode users who found Oh My OpenCode and want to know what it is now
- Written for
- OpenCode
- Time
- 9 min read · updated Nov 21, 2026
The short answer
Oh My OpenCode started as a plugin for OpenCode. The project is now called OmO and its repository is code-yeongyu/oh-my-openagent; the version the README leads with is a standalone omo command, not an OpenCode plugin. Its install line is curl -fsSL https://get.omo.dev/install.sh | bash, or bun add -g omo-ai and npm i -g omo-ai (the package is omo-ai, not omo). If you already use OpenCode, omo setup copies your provider keys, MCP servers and skills across without writing to OpenCode’s files. The license is the Sustainable Use License, not an open-source one: read it before using OmO for client work or passing it on. We read the README, docs and install script; we did not run it.
People searching for “oh my opencode” are usually looking for one GitHub project that has changed names twice. It was oh-my-opencode, then oh-my-openagent, and its README now calls it OmO. GitHub’s API still answers at the old name and returns code-yeongyu/oh-my-openagent. This page reads that repository’s README and docs, the install script it tells you to pipe into bash, and the npm registry, all on 21 November 2026. We did not run OmO, so it tells you what the files say and what to check, not whether it is any good.
What it is now
The README describes OmO as one omo command that takes a job and works on it with agents: research, code, a deck. It says OmO runs on senpi, its fork of an agent harness called pi. That is the change that matters for an OpenCode user. The project’s migration guide names two editions: the OpenCode plugin (oh-my-openagent 4.19.x, or a 5.x beta), and OmO Native, a standalone program installed from the omo-ai npm package or the install script, with “no host app, no plugin registration”. The README leads with the standalone one. The latest release in our check was v5.1.79 on 19 November 2026, and the last push was on 21 November.
So “Oh My OpenCode” can now mean three things: the old plugin you may already have, the standalone OmO, or OpenCode itself. The guide says you do not have to choose; the two editions can run side by side.
How to install it
To update, the README says to run the install line again, or omo update for a package-manager install. To remove it, rm ~/.local/bin/omo for the binary install, or bun remove -g omo-ai or npm uninstall -g omo-ai for the package.
What the install script does
Piping a script into bash runs whatever is at that address on the day. We downloaded install.sh (347 lines, 16 KB) and read it without running it. This is what it does:
- Downloads a native binary for your OS and CPU, plus a
SHA256SUMSfile, from the project’s mirror atget.omo.devand falls back to GitHub Releases if the mirror fails. It compares the binary’s checksum with that file and stops on a mismatch. Both files come from the same place, so the check catches a corrupted download, not a compromised source. - Writes the launcher to
~/.local/bin/omo, or the folder inOMO_INSTALL_DIR. - Edits your shell profile, adding a marked
PATHline to~/.zshrc,~/.bash_profile,~/.bashrcor~/.profile, unless you setOMO_NO_MODIFY_PATH=1. - Writes a receipt at
~/.omo/install.json. - Looks for other
omocommands on yourPATH, such as theomo-ai,oh-my-openagentoroh-my-opencodenpm packages, and asks before removing one. The default answer is no. With--remove-other-installsit removes them without asking, and in a non-interactive run it removes nothing. - Refuses to install into root’s home when started with
sudo, unlessOMO_INSTALL_ALLOW_SUDO=1. - Prints a telemetry notice: anonymous usage data, which it says excludes prompts, code, file paths and account data, goes to PostHog in the United States.
DO_NOT_TRACK=1orOMO_DISABLE_TELEMETRY=1turns it off.
The omo-ai npm package is different in one way: its package.json has a postinstall script, node bin/senpi-patch.mjs, which npm runs on install. We read that 34-line file in the package. It prepares the package’s launcher and the bundled engine on your disk through helper modules we did not read.
If you already use OpenCode
The README’s line for you is “Run omo setup once.” The migration guide says setup reads OpenCode’s own files (its auth.json and its config), prints one summary of everything it would carry over, asks one question, and writes only after you say yes. It writes to ~/.omo/, keeps a timestamped backup of each file before rewriting it, never overwrites an entry that already exists in omo, and never writes to OpenCode’s files. --dry-run prints the summary and stops. omo doctor then shows which task categories your providers can run.
The habits change too. Per the guide, OpenCode’s /models becomes /model, /connect becomes /login <provider>, and the Tab key that switches between build and plan has no equivalent, because omo has one main session and Shift+Tab cycles the thinking level instead. If an old OpenCode plugin of oh-my-openagent 4.19.x is still installed, its own omo command can occupy the name; omo --version printing a bare 4.19.4 means that. The first time omo starts, it also moves the old plugin’s config files into a backup folder under ~/.omo/, which a 4.19.x plugin will not find, and the guide gives the cp line to put it back.
How its skills differ from OpenCode’s
A skill in both is a folder with a SKILL.md. OpenCode’s docs say it looks in ~/.config/opencode/skills, ~/.claude/skills and ~/.agents/skills for you, and in .opencode/skills, .claude/skills and .agents/skills for a project. The OmO features reference lists project and user folders that include .opencode/skills, .claude/skills, .agents/skills and ~/.agents/skills, and says a same-named skill in a higher-priority folder overrides a lower one. So a skill folder you already use in OpenCode is in a place OmO also reads.
What OmO adds is its own set. Its docs list built-in skill sets such as debugging, frontend, git-master, playwright and visual-qa, and say they are activated by matching task intent rather than preloaded. They say a skill’s SKILL.md frontmatter can embed an MCP server, which plain OpenCode skills are not described as doing, and that disabled_skills in its config turns a built-in off. The setup command is stricter than OpenCode about one thing: it will not copy a skill folder with no top-level SKILL.md, or a SKILL.md with no description, because the engine would not load it.
The license
GitHub lists the license as “Other” (its API returns NOASSERTION). The LICENSE.md file is headed Sustainable Use License, version 1.0. In plain terms it says you may use or modify the software only for your own internal business purposes or for non-commercial or personal use, and that you may distribute it only free of charge for non-commercial purposes. That is not the same as an open-source license, and it is worth reading before you use OmO in client work or pass copies on. This is a plain reading, not legal advice. The npm packages oh-my-opencode and oh-my-openagent are still published too, under the identifier SUL-1.0.
Skills that make a finished file
OmO and OpenCode are agents: they decide what to do. A skill is what makes a finished file. Lookbk’s skills each render one, such as a landing page, a slide deck, a link-preview image or a title video, and each page shows what it made. They are plain skill folders, and OpenCode reads ~/.agents/skills, which is where the installer’s --shared flag writes. These four are free.
Start a new session and ask your agent which skills it can load. The OpenCode page lists the folders and what to do when a skill does not show up.
Common questions
- What is Oh My OpenCode?
- A project that began as a plugin for OpenCode and is now called OmO. Its repository is code-yeongyu/oh-my-openagent, and its README leads with a standalone
omocommand rather than a plugin. - How do I install Oh My OpenCode?
- The README’s line is
curl -fsSL https://get.omo.dev/install.sh | bash, orbun add -g omo-aiornpm i -g omo-ai. The package isomo-ai, notomo. Read the install script first. - Is Oh My OpenCode open source?
- The repository is public, but its license is the Sustainable Use License, which limits use to your own internal business purposes or non-commercial or personal use. That is not an open-source license. Read LICENSE.md in the repository.
- Does omo setup change my OpenCode config?
- The migration guide says setup never writes to OpenCode’s files. It reads them and writes to
~/.omo/, with a backup of each file it rewrites. The first omo session does move the old oh-my-openagent config files out of~/.config/opencode/into a backup folder. - Does OmO read the skills I already use in OpenCode?
- Its docs list the same project and user folders OpenCode reads, including
.agents/skillsand~/.agents/skills.omo setupalso copies global skills from OpenCode’s config folders into~/.omo/agent/skills/. - Does OmO collect data?
- The install script’s notice says it sends anonymous usage data to PostHog in the United States, and that
DO_NOT_TRACK=1orOMO_DISABLE_TELEMETRY=1turns it off. The README’s bug-report command lists what it would send and lets you read the report first. - How do I uninstall OmO?
rm ~/.local/bin/omofor the script install, orbun remove -g omo-aiornpm uninstall -g omo-aifor the package, per the README. The installer also added aPATHline to your shell profile, between lines marked as the omo installer, which you can delete.
Sources
What this guide says about other companies’ tools comes from their own documentation, checked on Nov 21, 2026.
Did this guide help?
Read next
All guidesCursor agent skills: add a skill from GitHub, step by step
How to get a skill from a GitHub repository into Cursor: the folders it loads, the plugin route in its docs, Cloud Agents and the frontmatter.
Safety checks before running a skill
A skill is instructions and code your agent may run. Six checks to make first: its source, its commands, its network use and what it may do without asking.
Install a skill in any agent
Install a Lookbk skill in Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI or OpenCode: where each looks, how to check it loaded and how to run it.
Where skills live: folders for every agent
The folders Claude Code, Codex, Cursor and other agents read skills from, which copy wins when two share a name, and what cloud agents can see.
Everything Claude Code (ECC): install it, and what to check first
The GitHub project once called everything-claude-code is now ECC. How its README installs it, what it adds to Claude Code, how to verify and undo it.
See a skill’s result before you install it.
Every skill page shows what it makes, what it needs and which agents can run it.



