# AI use cases for Harness Mix (emo-xiaoyu/harness-mix)

Repo: [emo-xiaoyu/harness-mix](https://github.com/emo-xiaoyu/harness-mix) | Use cases: 3 | Page: https://dedupedai.com/use-cases/emo-xiaoyu-harness-mix/

Harness Mix is a local kernel that adds a native Codex Desktop interface for connecting many coding agent harnesses and passing tasks between them.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Hand off a coding task between harnesses in one window: https://dedupedai.com/use-cases/emo-xiaoyu-harness-mix/hand-off-task-between-harnesses/
- Run a multi-harness agent team on a code review: https://dedupedai.com/use-cases/emo-xiaoyu-harness-mix/run-agent-team-review-in-codex/
- Install one skill or MCP server for many harnesses: https://dedupedai.com/use-cases/emo-xiaoyu-harness-mix/share-skills-and-mcp-across-harnesses/

## How to hand off a coding task between harnesses in one window with Harness Mix

Repo: [emo-xiaoyu/harness-mix](https://github.com/emo-xiaoyu/harness-mix) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Keep one Codex Desktop window and pass a task from one harness to another for a second pass.

**Who it is for:** A developer who has a half-finished change and wants a different harness to analyze it, implement it, or review it without leaving Codex Desktop.

**The problem:** Switching between separate coding agent apps loses context, uncommitted changes, and review notes.

**The outcome:** The task moves to another harness inside the same Codex Desktop window with its history, Git state, and review record kept in place.

### Steps

1. Install Node.js LTS from nodejs.org, then run `npm install --global @harness-mix/cli`.
2. Run `harness-mix` and wait for Codex Desktop to restart.
3. In Codex Desktop, open a task and choose one of the installed harnesses, then use `/switch` or the handoff badge in the input box.
4. Pick one of the four handoff modes: continue, run the previous plan, independent review, or re-analyze.
5. Check that the target harness shows the saved checkpoint and that its native session resumed.

### Prompt to paste into your agent

```text
Using the repo emo-xiaoyu/harness-mix, set up Harness Mix on this machine and hand the current task at <task description> from the current harness to <target harness> with the independent review mode. Report the checkpoint created and confirm the target harness resumed its native session.
```

### Caveat

Only harnesses that are installed locally and pass the handshake run for real, so the handoff target must be set up first.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/emo-xiaoyu-harness-mix/hand-off-task-between-harnesses/

## How to run a multi-harness agent team on a code review with Harness Mix

Repo: [emo-xiaoyu/harness-mix](https://github.com/emo-xiaoyu/harness-mix) | Difficulty: Advanced | Time: 1 hour | Basis: From the README

Start a team template so a lead harness coordinates named member harnesses on one goal.

**Who it is for:** A tech lead who wants several different harnesses to review, test, or refactor the same project in parallel and see the results in one place.

**The problem:** Parallel agent work usually comes back as separate outputs with no shared task list or way for members to talk.

**The outcome:** A named team with a lead, member roles, a shared task graph, and a native Team Workbench view that tracks progress.

### Steps

1. Run `npm install --global @harness-mix/cli` and then `harness-mix`.
2. Install and log in to at least two member harnesses, such as Codex and Claude Code, and confirm they show as ready in 设置 → 健康.
3. In the Codex Desktop input box, type `#` and open the 「团队」 page.
4. Choose the built-in 代码评审 template, assign a harness to each empty member slot, and type the review goal after the template.
5. Open the team dashboard from the top of the conversation and click 「展开详情」 to follow tasks and member messages.

### Prompt to paste into your agent

```text
Using the repo emo-xiaoyu/harness-mix, start the built-in 代码评审 agent team template for the project at <project path> with the goal <review goal>. Report each member harness, its assigned role, and the task state, and confirm that every member is using its own native session.
```

### Caveat

The team needs at least two harnesses installed and logged in on the machine, and the README marks some harnesses such as Kimi Code as still pending real-turn checks.

Risk, credentials: Each team member runs on its own harness account, so model usage and quota limits fall on each harness login rather than on Harness Mix.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/emo-xiaoyu-harness-mix/run-agent-team-review-in-codex/

## How to install one skill or MCP server for many harnesses with Harness Mix

Repo: [emo-xiaoyu/harness-mix](https://github.com/emo-xiaoyu/harness-mix) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Add a SKILL.md or remote MCP server once in Settings and use it across supported harnesses.

**Who it is for:** A developer who keeps the same team skills or internal MCP tools and wants every harness to see them without editing each harness's config by hand.

**The problem:** Each coding agent keeps its own skill folders and MCP settings, so the same setup has to be repeated for each tool.

**The outcome:** A skill installed through drag and drop, or a remote MCP server with custom headers, available to the selected harnesses.

### Steps

1. Run `npm install --global @harness-mix/cli`, then run `harness-mix` to open Codex Desktop with Harness Mix.
2. Open 设置 → Skills and drag a single `SKILL.md` file or a full skill folder into the drop area.
3. Choose Global or Project scope for the skill.
4. Open 设置 → MCP, add a Streamable HTTP or SSE server, and add any custom headers it needs.
5. Start a new session with a harness and confirm the skill or MCP tool appears in that harness.

### Prompt to paste into your agent

```text
Using the repo emo-xiaoyu/harness-mix, install the skill at <path to SKILL.md> at Global scope and register the remote MCP server at <server URL>. Report the install location for each harness and confirm that the skill is listed in one harness session.
```

### Caveat

Skill and MCP support differs by harness, and the matrix marks MCP as not declared for Pi and Oh My Pi.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/emo-xiaoyu-harness-mix/share-skills-and-mcp-across-harnesses/