
Steel Browser
This is a cloud browser automation service built for AI agents that need to handle JavaScript-heavy sites, login flows, and anti-bot measures. It wraps a real headless browser with a CLI that gives you both one-shot APIs (scrape, screenshot, PDF) and full session control for multi-step tasks. The session model is explicit: start a browser, navigate, snapshot to get element refs, interact using those refs, re-snapshot after DOM changes, then stop. It handles the cases where a simple HTTP fetch fails: pages that render client-side, forms that need filling, sites behind CAPTCHAs. The skill triggers when you need Claude to actually perform web tasks rather than write automation code for you to run later.
- skills.sh installs
- 1.8K
- GitHub stars
- 33
- last updated
- Aug 21, 2026
Install
$ npx -y skills add steel-dev/cli --skill steel-browser --agent claude-codeInstalls into .claude/skills of the current project.
Files
This is a cloud browser automation service built for AI agents that need to handle JavaScript-heavy sites, login flows, and anti-bot measures. It wraps a real headless browser with a CLI that gives you both one-shot APIs (scrape, screenshot, PDF) and full session control for multi-step tasks. The session model is explicit: start a browser, navigate, snapshot to get element refs, interact using those refs, re-snapshot after DOM changes, then stop. It handles the cases where a simple HTTP fetch fails: pages that render client-side, forms that need filling, sites behind CAPTCHAs. The skill triggers when you need Claude to actually perform web tasks rather than write automation code for you to run later.
Browse the repository on GitHubSponsors