Skill Detail

Control a real browser from MCP clients with Browser MCP

Use Browser MCP to let an MCP-capable agent drive a user's browser session for inspection, navigation, form work, and browser-state-aware automation.

Browser AutomationMCP
Browser Automation MCP Published
โญ 6.6k GitHub stars โฌ‡ 29.8k/wk npm
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill control-a-real-browser-from-mcp-clients-with-browser-mcp Copy
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Tools required
Browser MCP extension, @browsermcp/mcp server package, and an MCP-compatible client
Install & setup
Install the Browser MCP browser extension, add the MCP server to a compatible client using npx @browsermcp/mcp, then connect the extension to the browser session before asking the agent to inspect pages, navigate, fill forms, or capture browser-state evidence.
Author
Browser MCP
Publisher
Organization
Last updated
May 25, 2026
Quick brief

Use Browser MCP when an agent needs to operate against an actual browser session instead of a detached scraper, headless test runner, or static page fetch. The operator installs the Browser MCP extension, starts the MCP server, connects it to an MCP client such as Claude Desktop, Cursor, or another compatible host, and scopes the task to page navigation, visible-state inspection, form interactions, screenshots, tab context, and repeatable browser actions. Inputs are the running browser session, the MCP client configuration, target URLs, credentials already available to the user, and explicit task boundaries; outputs are completed browser actions, captured page state, evidence from the live session, and follow-up automation results. The scope boundary is live browser control through an MCP server and extension, not a generic browser automation framework, browser product, SDK, or web-scraping library listing.

How it works

What this skill actually does

Inputs and prerequisites: Browser MCP extension, @browsermcp/mcp server package, and an MCP-compatible client.

Setup notes: Install the Browser MCP browser extension, add the MCP server to a compatible client using npx @browsermcp/mcp, then connect the extension to the browser session before asking the agent to inspect pages, navigate, fill forms, or capture browser-state evidence.

Source and verification boundary: use https://browsermcp.io as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.

Framework fit: publish this as a MCP workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.