claude-foreman plugin
Claude Code supervises while Codex, opencode and pi work on your quota: delegated edits, independent review, model discussions
- 122 ★ on the repository
- Last commit
- MIT
Broad access
Why this level
- Delegated agents edit project files on their own on your subscription; verification happens on the diff after the run.
- Code and context go to the connected model providers.
Install
- Works in
Manual install
/plugin marketplace add DaydayupV/claude-foremanThen /plugin install foreman-en@claude-foreman; for the Chinese version foreman@claude-foreman.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- Delegated agents edit project files on their own on your subscription; verification happens on the diff after the run.
- Code and context go to the connected model providers.
This is third-party code. Review the repository files before installing.
What it does
claude-foreman is a plugin where Claude Code plans, decides and checks the result while handing hands-on coding to CLI agents such as Codex, opencode and pi according to your config. It also asks them for independent code review or gathers several models' views on a complex design before Claude Code decides. Claude Code reviews every returned diff, runs it and sends it back for rework if needed, and treats other models' views in a discussion as input only, accepting them on merit. Agents run on your Codex subscription, opencode plan or API credit and spend no Claude Code quota. Each handoff shows who got the task, on which model and effort, and a live view runs in the terminal and a web board.
Who it is for. People with subscriptions to several models who want Claude Code to supervise and verify work while other agents carry it out.
Good fit when
- Claude Code quota goes to routine edits and you want to move them to Codex.
- A complex architecture decision is worth discussing with several models before choosing.
- An independent diff review from another model is needed before acceptance.
- You want to see each agent's progress, speed and cost in one panel.
Not a fit when
- You have no Codex, opencode or pi: the plugin installs but Claude Code does all the work itself.
- You need screenshot handling by API models: they join discussions with text only.
- Policy forbids sending code to third-party model providers.
Example request
Hand this plan's function to Codex to implement, check the returned diff, run the tests and send it back for rework if anything is off.Limitations
Claude Code, Node.js 18 or newer and Git 2.25 or newer for tracking agent changes are required. Inside WSL Node.js must be installed in WSL itself or the plugin hooks fail before every command. Connect as many agents as you like: the Codex CLI is recommended and image generation works only through it. The Agy CLI is verified on Windows 11 only so far. Models with an OpenAI-compatible API join discussions and reviews with text only. To upgrade through an uploaded file the old version must be deleted separately. The latest commit is from November 2026.
How to disable. Remove the plugin with `/plugin` in Claude Code or from the My Uploads section in the Claude app. Configuration files and model keys in your setup are removed by hand.
FAQ
How do I install the plugin?
With `/plugin marketplace add DaydayupV/claude-foreman` and `/plugin install foreman-en@claude-foreman`. For the Chinese version install `foreman@claude-foreman`. In the Claude app you can upload the release file.
What does the setup command do?
`/foreman-en:setup` shows which tools the machine has, available models with effort levels and the current config: who gets each level of each task type.
Whose quota do the agents use?
Yours: a Codex plan, an opencode subscription or API credit. They do not spend Claude Code quota.
Does Claude Code accept others' output unchecked?
No. Claude Code reviews every diff, runs it and sends it back for rework when needed. It accepts or rejects other models' opinions in discussions on merit.
How do I watch the process live?
After each handoff Claude Code opens a live view in the app's terminal panel, and the optional web board shows the conversation map, per-seat progress, speed and cost.
Discussion
No questions or reviews yet.
Related
PluginsA Claude Code plugin that hands coding tasks to Codex agents, monitors them and verifies the results independently
Workflows for Claude Code: /do, /omo, /bmad-pilot and more commands, with Codex, Gemini or OpenCode writing the code through codeagent-wrapper
A Claude Code skill that hands implementation-heavy work to Codex CLI while Claude keeps planning, review and acceptance.
A plugin that syncs Claude Code, Codex CLI and Antigravity on one project and lets them hand work to each other.
A one-person coding factory: Claude Code manages the project while Codex, Qwen, Grok and Kimi write code as separate processes.
Multi-agent orchestration for Claude Code: teams, roles, autopilot and task routing across skills and subagents
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| claude-foreman (this card) | no | 122 | ||
| delegate-skills | no | 2,597 | general delegation skills; claude-foreman drives handoff, review and discussions across Codex, opencode and pi with a live board | |
| Codex Bridge | no | not required | 120 | a bridge between Claude Code and Codex; claude-foreman adds opencode, pi, model discussions and a web board |
README in short
The README covers installation through the Claude app and the marketplace, requirements, a breakdown of the configuration and five kinds of agents. Examples show adding API models, opencode, pi and agy.