cbrowser 9.3.1 → 9.4.1

package/README.md

@@ -5,8 +5,47 @@
 Built on Playwright with cognitive user simulation, constitutional safety boundaries, and the only UX testing that models how humans actually think.
 
 [![npm version](https://badge.fury.io/js/cbrowser.svg)](https://www.npmjs.com/package/cbrowser)
+[![Tests](https://github.com/alexandriashai/cbrowser/actions/workflows/test.yml/badge.svg)](https://github.com/alexandriashai/cbrowser/actions/workflows/test.yml)
 [![License: BSL-1.1](https://img.shields.io/badge/License-BSL--1.1-blue.svg)](LICENSE)
 [![MCP Ready](https://img.shields.io/badge/MCP-Remote%20%2B%20Local-blue)](https://modelcontextprotocol.io)
+[![TypeScript](https://img.shields.io/badge/TypeScript-5.3-blue.svg)](https://www.typescriptlang.org/)
+[![Node](https://img.shields.io/badge/Node-%3E%3D18-green.svg)](https://nodejs.org/)
+
+---
+
+## The Story Behind CBrowser
+
+I spent 16 years in international higher education marketing and UX, where I learned that the biggest accessibility and usability problems are the ones you can't see coming. A button that works perfectly for me might be invisible to a first-generation student in rural Indonesia using a slow connection on a small screen.
+
+Traditional testing tools couldn't help—they tested whether buttons clicked, not whether real humans could use them.
+
+Then came the AI agent revolution. I watched autonomous agents navigate websites, and I saw two problems:
+
+1. **They broke things.** Agents would accidentally submit forms, delete data, or get stuck in infinite loops. There were no guardrails.
+
+2. **They exposed how shallow our testing was.** If an AI couldn't figure out a UI, what chance did a confused first-timer have?
+
+I realized: **The industry needs tools that think like users, not like developers.** Tools that model patience, frustration, and the moment someone gives up. Tools that prevent autonomous agents from causing harm.
+
+So I built CBrowser.
+
+### Why These Technical Choices
+
+| Decision | Why |
+|----------|-----|
+| **TypeScript** | Type safety for a complex codebase; better IDE experience for contributors |
+| **Playwright over Puppeteer** | Cross-browser support (Chromium, Firefox, WebKit) out of the box; better auto-waiting |
+| **Constitutional AI safety** | AI agents need boundaries. Borrowed from Anthropic's constitutional AI principles to create risk zones |
+| **Cognitive simulation** | Based on UX research (Nielsen Norman Group, Baymard Institute) on how humans actually scan and interact with pages |
+| **MCP-first architecture** | Built for the Claude ecosystem. MCP is the future of AI tool integration |
+
+### What I Learned Building This
+
+- **Shipping beats perfecting.** 80+ npm versions taught me to iterate in public
+- **Real users reveal real problems.** Features I thought were clever often weren't useful; user feedback redirected the roadmap
+- **AI safety is a UX problem.** The same patterns that make sites confusing to humans make them dangerous for agents
+
+---
 
 ## Why CBrowser?
 
@@ -33,9 +72,9 @@ Most browser automation tests if buttons click. CBrowser tests if **real humans*
 
 ## Quick Start
 
-### Option 1: PAI Skill Installation (Claude Code Users)
+### Option 1: Claude Code Skill Installation
 
-If you use [Claude Code](https://claude.ai/claude-code) with [PAI (Personal AI Infrastructure)](https://github.com/danielmiessler/Personal_AI_Infrastructure), install CBrowser as a skill:
+If you use [Claude Code](https://claude.ai/claude-code), install CBrowser as a skill for seamless AI-assisted browser automation:
 
 ```bash
 # One-line installation
@@ -48,7 +87,7 @@ npx cbrowser install-skill
 This installs CBrowser to `~/.claude/skills/CBrowser/` with full skill structure:
 - `SKILL.md` - Main skill file with workflow routing
 - `Workflows/` - Navigate, Interact, Extract, Test, Journey workflows
-- `Tools/CBrowser.ts` - CLI wrapper for PAI
+- `Tools/CBrowser.ts` - CLI wrapper
 - `.memory/` - Session, selector, and persona storage
 
 After installation, add to your `~/.claude/skills/skill-index.json`:
@@ -919,18 +958,12 @@ Connect claude.ai directly to a remote CBrowser server:
 1. Deploy CBrowser on your server ([full guide](docs/REMOTE-MCP-SERVER.md))
 2. In claude.ai: Settings > Connectors > Add Custom Connector
 3. Add URL: `https://your-cbrowser-domain.com/mcp`
-4. Configure OAuth with Auth0 ([setup guide](docs/AUTH0-SETUP.md))
 
 **Public Demo (rate-limited):** `https://cbrowser-mcp-demo.wyldfyre.ai/mcp`
 - No authentication required
 - Rate limit: 5 requests/minute, burst of 10
 - For evaluation purposes only
 
-**Authenticated Server:** `https://cbrowser-mcp.wyldfyre.ai/mcp`
-- **OAuth 2.1 via Auth0** - For claude.ai web interface ([setup guide](docs/AUTH0-SETUP.md))
-- **API Key** - For Claude Code CLI and programmatic access
-- No rate limits for authenticated users
-
 ### Local MCP (Claude Desktop)
 
 Run CBrowser locally for Claude Desktop:
@@ -1161,7 +1194,7 @@ See the [`examples/`](examples/) directory:
 - [`basic-usage.ts`](examples/basic-usage.ts) - Navigation, extraction, sessions
 - [`smart-automation.ts`](examples/smart-automation.ts) - Smart click, assertions, test generation
 - [`visual-testing.ts`](examples/visual-testing.ts) - AI visual regression, cross-browser, responsive, A/B comparison
-- [`remote-mcp.ts`](examples/remote-mcp.ts) - Remote MCP server, Auth0 OAuth, demo server setup
+- [`remote-mcp.ts`](examples/remote-mcp.ts) - Remote MCP server setup and demo server connection
 - [`cognitive-journey.ts`](examples/cognitive-journey.ts) - Cognitive user simulation with personas, abandonment, and friction detection
 
 ### Workflow Recipes

package/dist/cli.js

@@ -1166,7 +1166,7 @@ Documentation: https://github.com/alexandriashai/cbrowser/wiki
     const daemonRunning = await isDaemonRunning();
     if (daemonRunning && ["navigate", "click", "fill", "hover", "screenshot", "extract", "run"].includes(command)) {
         console.log("🔌 Connected to running daemon");
-        let daemonCommand = command;
+        const daemonCommand = command;
         let daemonArgs = {};
         switch (command) {
             case "navigate":
@@ -3789,7 +3789,7 @@ Documentation: https://github.com/alexandriashai/cbrowser/wiki
                     const stepBrowser = new CBrowser({ headless });
                     await stepBrowser.launch();
                     try {
-                        let totalSteps = suite.tests.reduce((sum, t) => sum + t.steps.length, 0);
+                        const totalSteps = suite.tests.reduce((sum, t) => sum + t.steps.length, 0);
                         let stepNum = 0;
                         let stopped = false;
                         for (const test of suite.tests) {