BrowserAct Skills

BrowserAct Skills is a free, open-source repository of ready-to-use browser automation skills designed for OPENCLAW-style AI agents and skill workflows. It addresses the most common failure points in agent-driven browsing — Cloudflare blocks, CAPTCHA walls, session drops, and mid-flow action failures — by providing a powerful CLI engine with built-in stealth and captcha-solving capabilities. The project is MIT-licensed and works seamlessly across major AI coding platforms including Claude Code, Cursor, and OpenCode.

• Anti-Detection Stealth — Bypasses Cloudflare, reCAPTCHA, Datadome, and similar bot-detection systems using authentic browser fingerprints so agents are never blocked.
• Real Chrome Control — Drive your existing Chrome browser with all saved logins, cookies, and extensions, eliminating the need to re-authenticate.
• Parallel Execution — Run multiple stealth browsers concurrently, each with independent fingerprints, proxies, and sessions for high-throughput workflows.
• Built-in Captcha Solving — Automatically solves captchas without requiring third-party services or manual intervention.
• Low Token Noise — Strips up to 90% of junk HTML before feeding page content to the LLM, reducing cost and improving response speed.
• Proxy & Privacy Modes — Per-browser HTTP/SOCKS5 proxy support and privacy mode for fresh, isolated environments on every launch.
• Pre-built Skill Catalog — Includes ready-to-use skills for Amazon ASIN lookup, best-selling product extraction, Google News, Google Maps scraping, and YouTube transcript extraction.
• Cross-Platform Compatibility — Install skills with a single npx skills add command and use them immediately across all supported AI assistant platforms.
• Community & Credits — Star the repo and claim 500 free credits via Discord; request custom skills from the community.

To get started, install the skill with npx skills add browser-act/skills --skill browser-act and instruct your AI agent to perform any browser-based task. The CLI handles navigation, element interaction, screenshots, and session management from the command line.