# AI use cases for Open Pet Office (Gu-kai-lei/Open-Pet-Office) Repo: [Gu-kai-lei/Open-Pet-Office](https://github.com/Gu-kai-lei/Open-Pet-Office) | Use cases: 4 | Page: https://dedupedai.com/use-cases/gu-kai-lei-open-pet-office/ Open Pet Office is a Windows desktop app that shows Codex tasks as a visual multi-agent team, with a supervisor pet that plans, dispatches, and reviews work done by up to four worker agents in isolated workspaces. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Split a complex coding task across a supervised agent team: https://dedupedai.com/use-cases/gu-kai-lei-open-pet-office/run-multi-agent-mission-on-real-project/ - Track live Codex tasks from a floating desktop pet: https://dedupedai.com/use-cases/gu-kai-lei-open-pet-office/watch-live-codex-tasks-from-desktop-pet/ - Give each pet agent a DeepSeek or GLM model via OpenCodex: https://dedupedai.com/use-cases/gu-kai-lei-open-pet-office/route-deepseek-or-glm-models-to-pet-agents/ - Clone, test, and run Open Pet Office from source: https://dedupedai.com/use-cases/gu-kai-lei-open-pet-office/build-and-test-open-pet-office-from-source/ ## How to split a complex coding task across a supervised agent team with Open Pet Office Repo: [Gu-kai-lei/Open-Pet-Office](https://github.com/Gu-kai-lei/Open-Pet-Office) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Let a supervisor pet plan dependencies, dispatch up to four workers, and review results in isolated workspaces. **Who it is for:** A Windows developer with a real project and a task too large for one Codex chat, who wants the work split and checked. **The problem:** One long Codex session mixes planning, editing, and review, and parallel changes can collide in the main project. **The outcome:** A dependency plan, staged worker results, a supervisor review card, and changes written back to the main project only after review. ### Steps 1. Download `Pet-Office-*-portable.exe` from the Releases page and run it. 2. Confirm Codex CLI 0.155 or later is installed and signed in on this machine. 3. Hover over the supervisor pet, click the input icon, and turn on 分工 for the task. 4. Select the project, the Agent, and the model for each pet, then review and confirm the generated plan. 5. Watch the bell task center for approvals, questions, or conflicts until the final review appears on the task card. ### Prompt to paste into your agent ```text Using the Open-Pet-Office repo on Windows, start a 分工 Mission in the desktop app for the project at . Report the final supervisor review shown on the task card, and verify that a dependency plan was shown and confirmed before any worker ran. ``` ### Caveat The README says the real worker file-write acceptance is still blocked by a Codex Windows sandbox ACL issue in the current candidate build, so expect worker edits may not complete reliably. 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/gu-kai-lei-open-pet-office/run-multi-agent-mission-on-real-project/ ## How to track live Codex tasks from a floating desktop pet with Open Pet Office Repo: [Gu-kai-lei/Open-Pet-Office](https://github.com/Gu-kai-lei/Open-Pet-Office) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README See running, waiting, and finished Codex tasks on a pet card without hunting for the original window. **Who it is for:** A Codex user running several tasks at once who loses track of which chat needs attention. **The problem:** Codex tasks spread across windows, and a pending approval or question is easy to miss. **The outcome:** A task card under the pet showing the current model and progress, plus a bell task center grouping items that need action, in progress, and recently finished. ### Steps 1. Download `Pet-Office-*-portable.exe` from the Releases page and run it. 2. Confirm Codex CLI 0.155 or later is installed and signed in. 3. Start a Codex task in Codex Desktop; the supervisor pet shows a task card for it automatically. 4. Click the bell icon to open the task center and review the 需要处理 group first. 5. Click a task card to return to the matching Codex task. ### Prompt to paste into your agent ```text Using the Open-Pet-Office repo, run the Windows portable app while a Codex task is active in . Report what the pet task card shows for that task, and verify that the bell task center lists it under 进行中. ``` ### Caveat The session monitor reads local Codex JSONL logs, and the README warns that Codex upgrades may require a matching update to the app. 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/gu-kai-lei-open-pet-office/watch-live-codex-tasks-from-desktop-pet/ ## How to give each pet agent a DeepSeek or GLM model via OpenCodex with Open Pet Office Repo: [Gu-kai-lei/Open-Pet-Office](https://github.com/Gu-kai-lei/Open-Pet-Office) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Route third-party models into Codex and pick a different model for each pet in the team. **Who it is for:** A developer who wants cheaper or different models for some agents while keeping Codex as the session entry point. **The problem:** Codex by default uses OpenAI models, and a team may want a mix of providers for different roles. **The outcome:** Third-party models such as DeepSeek or GLM appear in the Codex model list and can be assigned to individual pets. ### Steps 1. Install the optional OpenCodex from https://github.com/lidge-jun/opencodex. 2. Add a DeepSeek or GLM provider in OpenCodex by following the OpenCodex documentation. (inferred, check the README) 3. Open Open Pet Office and check that the new model appears in the model list. 4. Select a different model for each worker pet, then keep the supervisor on the model you trust for planning. 5. Check the model quota display, which shows a provider balance only when the provider exposes it. ### Prompt to paste into your agent ```text For the Open-Pet-Office repo, configure OpenCodex with a DeepSeek or GLM provider, then select that model for one worker pet in . Report the chosen model name, and verify it appears in the pet's model selector. ``` ### Caveat The README gives no OpenCodex setup steps beyond linking to its repo, so the provider configuration must follow OpenCodex's own documentation. Risk, credentials: Provider API keys for DeepSeek, GLM, or other services must stay out of screenshots, logs, and issue reports, even though the app masks them on the desktop. 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/gu-kai-lei-open-pet-office/route-deepseek-or-glm-models-to-pet-agents/ ## How to clone, test, and run Open Pet Office from source with Open Pet Office Repo: [Gu-kai-lei/Open-Pet-Office](https://github.com/Gu-kai-lei/Open-Pet-Office) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Get a working development build and run the test suite before changing the app. **Who it is for:** A contributor who wants to fix a bug or try a feature in Open Pet Office on Windows. **The problem:** Changes to the app are risky without first confirming the existing tests pass on the local machine. **The outcome:** A running development build and a test result that shows whether the local environment matches the project. ### Steps 1. `git clone https://github.com/Gu-kai-lei/Open-Pet-Office.git` 2. `cd Open-Pet-Office` 3. `npm install` 4. `npm test` 5. `npm start` ### Prompt to paste into your agent ```text Clone https://github.com/Gu-kai-lei/Open-Pet-Office into , run `npm install` and then `npm test`. Report the test outcome, and verify that the run finishes without installation errors before you attempt `npm start`. ``` ### Caveat The README says the app is an early Windows preview whose interface, protocols, and data structures are still changing, so local source builds may break between updates. 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/gu-kai-lei-open-pet-office/build-and-test-open-pet-office-from-source/