dsh-computer-use
Run the following command in DeepSeek Harness:
dsh plugin install xiaoheizi1212/dsh-computer-use
Paste the following prompt into your AI chat to install this plugin:
In DeepSeek Harness, you can install this plugin by running the installation command: dsh plugin install xiaoheizi1212/dsh-computer-use. Source: https://github.com/xiaoheizi1212/dsh-computer-use
About this plugin
The dsh-computer-use plugin for DeepSeek Harness addresses the challenge of bridging the gap between AI agents and real-world computing environments. It provides a model-agnostic interface that enables Large Language Models like DeepSeek to control computers through text commands, effectively extending the capabilities of text-only agents.
This plugin offers a comprehensive set of core capabilities, including an isolated browser, a Windows native helper, and a visual perception layer. It supports multiple providers such as Playwright for cross-platform web automation and a dedicated Windows helper for desktop control. Additionally, features like the Chrome Cookie Bridge allow for importing logged-in sessions, while vision models enable the AI to analyze screenshots and accessibility trees for true UI understanding.
With operations involving real environments, security and control are paramount. The plugin includes a risk-classification system that determines approval requirements based on action type, alongside session logging and domain allowlists. This makes it an ideal solution for developers, QA engineers, and power users looking to implement robust, safe automation workflows using DeepSeek for computer control.
Use Cases
- Automate web testing tasks
- Import logged-in cookies for automation
- Analyze interface elements with vision models
Best For
- DeepSeek Harness developers
- QA engineers needing automation
- Researchers studying multimodal agent interaction
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.