engineering / skill
BrowserSkill
A command line tool, browser extension and agent skill that let an AI agent work in your signed-in Chrome or Microsoft Edge browser.
At a glance
- Purpose
- A command line tool, browser extension and agent skill that let an AI agent work in your signed-in Chrome or Microsoft Edge browser.
- Runs in
- Local agent with shell
- Operating systems
- macOS, Linux, Windows
- Needs
- Runtime: bsk CLI and browser extension; Needs internet: Yes; Needs accounts or API keys beyond your agent: Not stated
- Files and data
- Reads files: Yes; Writes files: Yes; Sends data to other services: Yes
- Tool type
- Skill
- Delivery
- From GitHub
- Version
- Latest on the main branch
- Access
- Code on GitHub
- Compatibility
- According to the README, the bsk CLI works with Cursor, Claude Code, Codex, OpenClaw, CodeBuddy, WorkBuddy, Pi, Hermes Agent or another shell-capable agent, and DeepSeek Harness has a dedicated plugin. The README lists CLI and daemon support for macOS, Linux and Windows. The README lists extension support for Chrome and Microsoft Edge based on Chromium 125 or later.
- External costs
- The repository is MIT-licensed and states no pricing. The README says BrowserSkill does not operate a mandatory cloud service. Your coding agent or model provider may charge for usage.
- Permissions
- The skill uses your coding agent's granted access. The README asks the agent to fetch and follow AGENT_INSTALL.md. Manual routes execute downloaded shell or PowerShell scripts, install under ~/.local/bin by default, and add a browser extension and agent skill. The README says the background daemon's Agent Window shares the selected profile's logins and website permissions. Tab borrowing asks by default; a browser setting disables confirmation. The README describes page reading, forms, browser evidence, local uploads/downloads and session request replay that can change server data. The README says a bsk background daemon checks releases every 30 minutes and installs when no agent session is active. Foreground daemons and installs lacking executable-directory write access report updates instead; BSK_AUTO_UPDATE=off disables daemon auto-updates.
- Usage terms
- MIT License, as stated in the repository. See its LICENSE file for the exact terms.
Code on GitHub
Tencent/BrowserSkill
MITLet AI agents use your real, logged-in browser without interrupting your work. CLI + extension for browser automation across any shell-capable AI agent.
Topics: agent browser-use dsh-plugin
Public GitHub data fetched · 0 days old
Skill folders in this repository (2)
Folders that contain a SKILL.md file, from GitHub's file list on .
What it does
BrowserSkill is a command-line tool, browser extension and agent skill. The README says it connects an AI agent to Chrome or Microsoft Edge using existing signed-in accounts. The README describes a bsk CLI with a background daemon, an extension and a skill for using the tool.The README says tasks run in a visible Agent Window and existing tabs can be borrowed explicitly. The README lists page reading, clicking, typing, tab management, screenshots, and local-mode file uploads and downloads. The README describes debugging capture for requests, response bodies, console output and page changes, together with request modification, blocking, mock responses and same-origin replay.
The README names Cursor, Claude Code, Codex, OpenClaw, CodeBuddy, WorkBuddy, Pi and Hermes Agent as CLI users, and describes a separate DeepSeek Harness plugin. The README also describes pairing a local browser with an agent running on a server.
Who it is for
People who want a shell-capable AI agent to read pages, fill forms and debug websites in a browser they are already signed into.Example tasks
Open a web page, summarize it and end the browser session.Fill a form or work through a website using your logged-in accounts.
Capture a full-page screenshot of a page.
Investigate why saving a form on a local site fails by capturing the request, response and console output.
Pair an agent running on a server with a browser on your computer.
Getting started
The README's quick-start route is to ask your agent to set up browser-skill by following the AGENT_INSTALL.md guide, then install the extension from the Chrome or Edge store. Manual setup has four steps: install the CLI with an install script (a shell script on macOS and Linux, a PowerShell script on Windows), connect the extension, run bsk install-skill, and check with bsk doctor. DeepSeek Harness users add a plugin instead, with dsh plugin --profile web add @wxg-prc-cpg/browser-skill-dsh-plugin. The README has the exact commands.Limitations
The README says local automation requires an AI agent, the bsk CLI and browser extension. The README lists CLI and daemon support for macOS on Apple Silicon and Intel, Linux on x64 and ARM64, and Windows x64, with Windows 10 or Windows Server 2016 or later. The README lists Chrome and Microsoft Edge based on Chromium 125 or later and says other Chromium browsers have no guaranteed compatibility. The README's DSH route requires DeepSeek Harness, pnpm and bsk with the extension connected.The README says an Agent Window shares the profile's login state and is not a separate account or security sandbox. The README says remote mode does not support file uploads or downloads, request replay can change server data, and debugging redaction can leave sensitive data. The README describes limitations for nested scrolling panels and virtualized lists in full-page captures. The README says component releases should match and store builds can lag behind the repository. The repository can change at any time; check its README before you install.