AI Agent 自主浏览器操作与反检测方案

You are an expert in browser automation for AI agents. Help me design a robust browser automation strategy that handles anti-bot detection. Cover the following: ## 1. Browser Environment Setup - Headless vs headed browser selection (Puppeteer/Playwright/custom Chromium) - Fingerprint randomization: canvas, WebGL, AudioContext, fonts, screen resolution - User-Agent rotation and consistency checks - WebDriver detection bypass (navigator.webdriver, CDP detection) ## 2. Behavioral Simulation - Human-like mouse movements (Bezier curves, random delays) - Scroll patterns (variable speed, pause at content) - Typing simulation (variable keystroke intervals, typos + corrections) - Tab switching and multi-tab browsing patterns ## 3. Network Layer - Proxy rotation strategy (residential vs datacenter) - Cookie and session management - TLS fingerprint consistency - Rate limiting and request throttling ## 4. CAPTCHA Handling - reCAPTCHA v2/v3 strategy differences - hCaptcha and Turnstile approaches - When to use solving services vs behavioral bypass ## 5. Agent Integration - How to expose browser actions as tool calls for LLM agents - Error recovery and retry logic - Screenshot-based visual verification loops For my specific use case: [describe your target site and automation goal] Provide concrete code snippets in Python/Node.js and a decision tree for choosing the right approach.