DSH Hub
dsh-tabbit cover

Tabbit-Browser/dsh-tabbit

dsh-tabbit

BundleWorkflow96 GitHub stars· updated 2026-08-22

Tabbit Browser plugins for Deepseek Harness

Install

npx @deepseek-ai/dsh plugin --profile web add dsh-tabbit

Restart `dsh web` after install. Bundle APIs can change during the developer preview.

README badge

dsh-tabbit DSH Hub badge
[![DSH Hub](https://dshhub.dev/badge/dsh-tabbit.svg)](https://dshhub.dev/plugins/dsh-tabbit)

Paste this into your README. The star count updates with every catalog sync.

From the README

Excerpt from Tabbit-Browser/dsh-tabbit, cleaned of badges and images.

tabbit-browser for DeepSeek Harness

English | 简体中文

A plugin for DeepSeek Harness (DSH) that gives the agent control over your Tabbit Browser: real pages, real login state, and real interactions, driven through tabbit-cli — the task-isolated Playwright CLI owned by the browser itself. Use it for web automation, information extraction, QA, and benchmarks.

What you get

ComponentDescription
tabbit-browser skillThe working guide for browser automation: persistent task spaces, locators and waits, screenshots, receipts and recovery. Discovered and loaded automatically with the plugin — no separate skill install. The model loads it via skill({ name: "tabbit-browser" }) or /tabbit-browser.
tabbit_browser_install toolEnvironment preflight: detects installed stable Tabbit editions, requires version 1.9.0 or newer, and verifies the tabbit-cli resident runtime. When Tabbit is missing or outdated, it starts a DSH background job that downloads the region-appropriate installer.
tabbit_plugin_update toolPlugin update check: compares the installed plugin version with the published changelog at most once a day, silently skips offline failures, and records a version the user declined. When a newer release exists, the skill loads with an update notice showing what the new version added.

Installation

1. Check or install DeepSeek Harness

Check whether DSH is already installed:

dsh --version

If the command prints a version number, continue to the next step. If it is not found, install it for your operating system.

macOS

Install Node.js 20 or newer, then install DSH:

brew install node
npm install -g @deepseek-ai/dsh

Windows

Install Node.js LTS in PowerShell:

winget install OpenJS.NodeJS.LTS

Reopen PowerShell after the installation, then install DSH:

npm install -g @deepseek-ai/dsh

Run dsh --version again to confirm DSH works.

2. Install the tabbit-browser plugin

dsh plugin --profile web add dsh-tabbit

This installs the npm package (prebuilt tarball, CDN-served). If the npm registry is unreachable, fall back to the GitHub source:

dsh plugin --profile web add github:Tabbit-Browser/dsh-tabbit

3. Start DSH

dsh web

How it works

After installation, the bundle automatically registers its skill provider. The model loads the skill via skill({ name: "tabbit-browser" }) or /tabbit-browser. Before the first browser operation in a task, the skill calls tabbit_browser_install:

Related plugins