AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
MCP verified Apache-2.0 Self-run

User SWeb Agent MCP

mcp-vsrupeshkumar-user-sweb-agent-mcp · by vsrupeshkumar

An AI-native browser control layer that turns your existing web session into a programmable interface—enabling agents to navigate, interact, and complete real-world tasks directly inside your browser with full context and zero friction.

No reviews yet
0 installs
18 views
0.0% view→install

Install

$ agentstack add mcp-vsrupeshkumar-user-sweb-agent-mcp

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access Used
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/mcp-vsrupeshkumar-user-sweb-agent-mcp)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
3mo ago

Declared compatibility

Claude CodeClaude DesktopCursorWindsurf

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of User SWeb Agent MCP? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Browser MCP: Autonomous Web Agent Kernel

[](https://github.com/vsrupeshkumar/User-sWeb-Agent---MCP) [](https://opensource.org/licenses/MIT)

Browser MCP is a high-performance Model Context Protocol (MCP) server designed to bridge the gap between Large Language Models (LLMs) and real-time browser environments. It facilitates secure, stealthy, and stateful browser automation by leveraging existing user profiles.


🏗 System Architecture

The following diagram illustrates the topological data flow between the LLM client, the MCP Kernel, and the Browser Runtime.

graph TD
    subgraph "AI Client (Claude/Cursor/VSCode)"
        Client[MCP Client]
    end

    subgraph "MCP Server Kernel"
        Server[MCP Server]
        Context[Context Manager]
        WS_Server[WebSocket Server]
    end

    subgraph "Browser Environment"
        Ext[Browser Extension]
        DOM[DOM Tree/Execution Context]
    end

    Client -- "JSON-RPC (stdio)" --> Server
    Server -- "Tool Call" --> Context
    Context -- "Binary/JSON" --> WS_Server
    WS_Server -- "Secure WebSocket" --> Ext
    Ext -- "CDP / Web API" --> DOM
    DOM -- "Mutation Events" --> Ext
    Ext -- "Response" --> WS_Server
    WS_Server -- "Result" --> Server
    Server -- "Final Output" --> Client

🔢 Formal Logic & Mathematical Foundation

1. State Transition Model

We define the browser automation process as a sequence of state transitions within a finite state machine (FSM).

Let $\mathcal{S}$ be the set of all possible browser states (DOM structure, cookies, local storage, etc.). Let $\mathcal{T}$ be the set of tools defined in the MCP schema. Let $\mathcal{A}$ be the domain of input arguments for these tools.

A tool execution is defined by the transition function $\delta$: $$\delta: \mathcal{S} \times \mathcal{T} \times \mathcal{A} \rightarrow \mathcal{S} \times \mathcal{R}$$

Where:

  • $s_t \in \mathcal{S}$ is the current state.
  • $t \in \mathcal{T}$ is the selected tool.
  • $a \in \mathcal{A}$ is the argument vector.
  • $r \in \mathcal{R}$ is the observation (text, image, or structured data) returned to the LLM.

2. Contextual Persistence

The MCP Server maintains a singleton context $\mathcal{C}$ that encapsulates the active WebSocket connection $\omega$: $$\mathcal{C} = \{ \omega \mid \omega \in \text{WebSocket}, \text{status}(\omega) = \text{'connected'} \}$$

The execution of any tool $t$ is conditional on the existence of a valid $\omega$: $$P(\text{Success} \mid t, a) = \begin{cases} 1 & \text{if } \omega \in \mathcal{C} \\ 0 & \text{if } \omega \notin \mathcal{C} \end{cases}$$


🚀 Core Features

  • ⚡ Zero Latency Core: Optimized for local execution, minimizing network overhead between the LLM and the target DOM.
  • 🛡 Stealth Protocol: Unlike traditional headless drivers, Browser MCP utilizes the user's authentic browser fingerprint, bypassing advanced anti-bot mechanisms and CAPTCHAs.
  • 🔒 Privacy First Architecture: All automation logic and data handling occur locally. No sensitive browser data is ever transmitted to external third-party servers.
  • 🔄 State Synchronization: Real-time synchronization between the browser extension and the MCP server via low-latency WebSockets.

🛠 Installation & Setup

Prerequisites

Build from Source

# Clone the repository
git clone https://github.com/vsrupeshkumar/User-sWeb-Agent---MCP.git

# Install dependencies
npm install

# Compile the kernel
npm run build

Configuration

Add the server to your MCP client configuration (e.g., claude_desktop_config.json):

{
  "mcpServers": {
    "browser-mcp": {
      "command": "node",
      "args": ["/path/to/dist/index.js"]
    }
  }
}

📜 License

This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details.


Built with ❤️ by vsrupeshkumar

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.