Workbuddy Local Worker
# WorkBuddy Bridge
[English](README.md) | [简体中文](README.zh-CN.md)
**Use WorkBuddy as a subagent without leaving your AI assistant.**
Delegate Origin plotting, Blender modeling, and other tasks from your current chat. Ask your main assistant to pass along the relevant context and use the same project folder, then bring WorkBuddy's results back into your main workflow.
**The bridge is free and needs no separate API key.** WorkBuddy's own account limits and any model charges still apply.
## What can you do?
- **Stay in one chat.** Your main assistant delegates tasks and collects the results.
- **Share context and a workspace.** Pass the task brief, decisions, and file paths so WorkBuddy can work with the same local project files.
- **Use WorkBuddy's tools.** Plot in Origin, model in Blender, or use other MCP tools and skills you have configured.
## Get started
Copy this and send it to your AI assistant to install:
> Install WorkBuddy Bridge for me: https://github.com/P-PPPP/workbuddy-bridge
[Manual installation](docs/REFERENCE.md#installation)
## Use it on your current project
Tell your AI assistant:
> Delegate this task to WorkBuddy. Share the relevant context from our conversation and the path to our current project folder. Ask me which model to use first, then save the results in that folder and report back here.
Your assistant chooses what context to send; the full chat history is not synced automatically.
## Example: make a chart in Origin
If Origin MCP is already set up and working in WorkBuddy, ask:
> Have WorkBuddy use Origin MCP to plot the data we are discussing. Pass along the data file paths, labels, and style requirements from this conversation. Ask me which model to use, then save the chart as PNG and PDF in our current project folder.
The Origin workflow was tested with OriginPro 2024 using synthetic sine and cosine data. WorkBuddy created the chart below and exported PNG and PDF files:
## Example: make a model in Blender
If Blender MCP is set up in WorkBuddy, try:
> Have WorkBuddy use Blender MCP to create a simple desk organizer. Pass along the dimensions and design requirements we agreed on. Ask me which model to use, then save the .blend file and a preview image in our current project folder and return the file paths here.
The Blender example is a suggested workflow; it has not been tested in this project.
You can also ask:
> Which MCP tools and skills can WorkBuddy use?
> Check the progress of my WorkBuddy task.
> Stop the current WorkBuddy task.
## A few things to know
- **Choose your model.** If you don't, WorkBuddy uses its own default, which may be Auto. Tasks use your WorkBuddy account and normal usage allowance.
- **These are separate task sessions.** They don't take over an existing desktop chat, and you shouldn't expect them to appear in WorkBuddy's chat history.
- **WorkBuddy's tools still need to be installed.** This project connects to your setup; it doesn't install Origin or other applications for you.
- **Keep local data private.** Task messages and results are saved on your computer. Don't upload your `data` folder or account files.
## Need help?
**Your AI app can't connect?** Check that Node.js 24+ is installed and that you copied the generated settings. From the project folder, run `npm run smoke` to check the connection without calling a model.
**WorkBuddy needs you to sign in?** Sign in again in the desktop app, then ask your assistant to start a new task session.
For tool details, custom settings, and troubleshooting, see the [technical reference](docs/REFERENCE.md). You can also [report an issue](https://github.com/P-PPPP/workbuddy-bridge/issues).
Tested with WorkBuddy 5.9.3 on Windows. This is an unofficial project; compatibility may change after WorkBuddy updates.
## License
[MIT](LICENSE). Personal and non-commercial use is encouraged; commercial use remains permitted.
TDQS
Scored across 7 tools
Most tools have clearly distinct roles: send_message starts or continues work, stop_task cancels, and the select_* tools configure model or MCP/skills. However, get_message, wait_for_message, and get_status overlap somewhat in observing task events and pending input, which could require careful reading of descriptions to choose correctly.
All seven tools follow a consistent snake_case verb_noun pattern such as send_message, get_status, stop_task, and select_model. The naming is predictable and easy to map to actions.
Seven tools is well-scoped for an asynchronous worker interface covering task messaging, observation, cancellation, and configuration. Each tool appears to earn its place without obvious redundancy.
The surface covers the full interaction lifecycle: starting/continuing tasks, reading incremental events, waiting for events or input, checking health/status, cancelling tasks, and selecting runtime model and MCP/skill configuration. No obvious operational gaps are apparent for a local async worker.