Setup

Install the semantic browser layer and connect your first agent.

Start locally, validate the browser scanner, then expose Sentinel as MCP tools to Claude Desktop, Claude Code, Cursor, or your own agent runtime.

Install runtime

Node 20+, npm, Playwright browser dependencies, and the Sentinel package.

Configure models

Choose Claude, OpenAI/ChatGPT, Mistral, OpenRouter, Ollama, or a custom compatible endpoint.

Expose tools

Run the MCP server or AgentBridge so agents can observe and act through semantic menus.

macOS

$ brew install node$ git clone https://github.com/Winzyss-code/sentinel-sfa.git$ cd sentinel-sfa$ npm install$ npx playwright install chromium$ cp .env.example .env$ npm run test:context

Linux

Ubuntu / Debian

$ sudo apt update$ sudo apt install -y nodejs npm git$ git clone https://github.com/Winzyss-code/sentinel-sfa.git$ cd sentinel-sfa$ npm install$ npx playwright install --with-deps chromium$ cp .env.example .env$ npm run test:context

Windows

Windows PowerShell

> winget install OpenJS.NodeJS> winget install Git.Git> git clone https://github.com/Winzyss-code/sentinel-sfa.git> cd sentinel-sfa> npm install> npx playwright install chromium> copy .env.example .env> npm run test:context

CLI

Run as MCP server

$ mkdir -p ~/.sentinel-sfa/browser-profile$ BROWSER_PROFILE_PATH=~/.sentinel-sfa/browser-profile npm run mcp[sentinel-sfa] MCP server ready on stdio# Then call from Claude Desktop:browser_open({ "url": "https://example.com" })browser_act({ "index": 1, "action": "click" })

Provider quick matrix

Provider	Use for	Env
Claude Code	agent planning + code workflows	ANTHROPIC_API_KEY, LLM_MODEL
Claude Desktop MCP	browser tools inside Claude	BROWSER_PROFILE_PATH, HEADLESS
OpenAI / ChatGPT	tool-calling agent runtime	OPENAI_API_KEY, OPENAI_MODEL
Mistral	fast cheaper model loop	OPENAI_BASE_URL, OPENAI_API_KEY
OpenCLAW / custom	external agent bridge	AgentBridge / MCP stdio