# AI use cases for Codex on Crack (ethanplusai/codex-on-crack)

Repo: [ethanplusai/codex-on-crack](https://github.com/ethanplusai/codex-on-crack) | Use cases: 4 | Page: https://dedupedai.com/use-cases/ethanplusai-codex-on-crack/

Codex on Crack is a JavaScript package that coordinates your configured AI models from inside Codex, handling planning, delegation, host tool handoffs, and review across workspaces.

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

- Plan a project and agree on which model does each part: https://dedupedai.com/use-cases/ethanplusai-codex-on-crack/plan-project-and-assign-models/
- Have Opus plan and a native worker build, then review: https://dedupedai.com/use-cases/ethanplusai-codex-on-crack/opus-plans-native-worker-builds/
- Build with Opus through Claude Code, then check in Codex: https://dedupedai.com/use-cases/ethanplusai-codex-on-crack/build-with-opus-then-check-in-codex/
- Follow and review agent work in the Codex on Crack panel: https://dedupedai.com/use-cases/ethanplusai-codex-on-crack/review-agent-work-in-panel/

## How to plan a project and agree on which model does each part with Codex on Crack

Repo: [ethanplusai/codex-on-crack](https://github.com/ethanplusai/codex-on-crack) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Get a plan, a model recommendation for each part, and your approval before any building starts.

**Who it is for:** A developer starting a new project in Codex who has several models available and is not sure which one should do which work.

**The problem:** Choosing models and splitting work across them by guesswork wastes time and can send the wrong task to the wrong model.

**The outcome:** An agreed plan that assigns each phase to a model based on your configured roles and declared capabilities, with nothing built until you approve it.

### Steps

1. Run `codex plugin marketplace add "$PWD"` from this checkout in a stable folder.
2. Run `codex plugin add codex-on-crack@codex-on-crack`.
3. Start a fresh Codex conversation and ask it to help plan the project, check which models are available, and recommend who should do what.
4. Review the proposed plan and the model recommendations, then agree on the plan before asking for implementation.

### Prompt to paste into your agent

```text
Using the codex-on-crack package in <project folder>, run the crack-plan skill to list my available models and propose a plan that assigns each phase to a model. Expected output: a written plan with one model recommendation per phase and no code changes. Verify that the plan waits for my approval before any build step begins.
```

### Caveat

Recommendations depend on the roles and capabilities you configure, and the package does not benchmark every model.

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/ethanplusai-codex-on-crack/plan-project-and-assign-models/

## How to have Opus plan and a native worker build, then review with Codex on Crack

Repo: [ethanplusai/codex-on-crack](https://github.com/ethanplusai/codex-on-crack) | Difficulty: Advanced | Time: 1 hour | Basis: From the README

Split one assignment so Opus plans and reviews while a configured worker implements it.

**Who it is for:** A developer with an Opus route and a compatible native worker who wants a stronger planner and a cheaper implementer on the same build.

**The problem:** Using one model for both planning and implementation may not fit the work, and manually moving context between models is error-prone.

**The outcome:** Opus produces the plan, Codex dispatches the implementation to the native worker, and the result returns to Opus and Codex for review.

### Steps

1. Run `codex plugin marketplace add "$PWD"` from this checkout in a stable folder.
2. Run `codex plugin add codex-on-crack@codex-on-crack`.
3. Read the setup guide and configure a compatible native worker with the crack-setup skill, reviewing the preview before applying it.
4. Start a fresh conversation and ask Opus to plan the feature and Flash to implement it, then bring the result back for review.
5. Check the returned worker results and host-tool evidence before accepting the work.

### Prompt to paste into your agent

```text
Use the codex-on-crack package in <project folder> to have Opus plan <feature name> and a configured native worker implement it, then return the result to Codex for review. Expected output: a plan, the worker's changes, and a review note from the lead. Verify that the review cites the worker results and host-tool evidence.
```

### Caveat

Instructions and file scopes are not an OS sandbox, so the worker can still act within your permissions.

Risk, credentials: Claude mode uses your authenticated Claude Code client, so keep your subscription login and any worker route credentials under your own control.

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/ethanplusai-codex-on-crack/opus-plans-native-worker-builds/

## How to build with Opus through Claude Code, then check in Codex with Codex on Crack

Repo: [ethanplusai/codex-on-crack](https://github.com/ethanplusai/codex-on-crack) | Difficulty: Advanced | Time: 1 hour | Basis: From the README

Let Opus implement a whole assignment, then have the Codex host verify it in the browser.

**Who it is for:** A developer who prefers Claude for a build but wants Codex tools, such as browser checks or computer tools, to verify the result.

**The problem:** External Claude sessions do not inherit Codex tools, so checking the finished build usually means a manual handoff.

**The outcome:** Opus implements the assignment through Claude Code, and the Codex host runs the requested browser or computer checks and reports the evidence back.

### Steps

1. Install the Claude Code CLI and confirm it is authenticated through your subscription.
2. Run `codex plugin marketplace add "$PWD"` from this checkout in a stable folder.
3. Run `codex plugin add codex-on-crack@codex-on-crack`.
4. Start a fresh conversation and ask Codex to have Opus implement the assignment through Claude Code.
5. Ask the Codex host to check the result in the browser and report what it found.

### Prompt to paste into your agent

```text
With the codex-on-crack package in <project folder>, have Opus 5.5 implement <assignment> through Claude Code, then ask the Codex host to open the result in the browser and report any problems. Expected output: the implementation plus a browser check summary. Verify that the browser evidence comes from the Codex host, not from the Claude session.
```

### Caveat

The current Claude adapter targets claude-opus-5-5, and other Claude models are not automatically supported.

Risk, credentials: The Claude Code CLI uses your own subscription login, so you are responsible for keeping that authentication secure.

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/ethanplusai-codex-on-crack/build-with-opus-then-check-in-codex/

## How to follow and review agent work in the Codex on Crack panel with Codex on Crack

Repo: [ethanplusai/codex-on-crack](https://github.com/ethanplusai/codex-on-crack) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Watch phases, handoffs, and tool activity, then leave feedback without interrupting the agent.

**Who it is for:** A developer running a multi-model build who wants to see progress and review screenshots without stopping the conversation.

**The problem:** Multi-agent work is hard to follow, and feedback sent mid-run can be lost or mistaken for approval.

**The outcome:** A panel view of phases, handoffs, timing, and token counters, plus saved feedback the agent reads at a checkpoint and an exportable offline replay.

### Steps

1. Run `codex plugin add codex-on-crack-panel@codex-on-crack` to add the panel.
2. Start a fresh conversation and ask Codex to connect this workspace and its runs to the Codex on Crack panel.
3. Open the Overview, Review, or Usage and routes view for your chosen workspace.
4. Leave feedback on a screenshot, then send an explicit approval only when you have reviewed the work.
5. Export an offline replay for later inspection.

### Prompt to paste into your agent

```text
Connect the workspace at <workspace path> to the Codex on Crack panel and list its registered phases, agents, and handoffs. Expected output: a summary of the run with timing and available token counters, with missing values shown as unknown. Verify that no feedback has been treated as approval.
```

### Caveat

The panel only shows sessions and runs you register, and it does not read every conversation on your computer.

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/ethanplusai-codex-on-crack/review-agent-work-in-panel/