ego-browser
ego-browser
ego-browser gives AI agents a CLI-accessible Node.js runtime, with built-in helpers - snapshotText, click, js, cdp, and more - that agents call directly inside JS scripts to observe pages, interact with UI, evaluate browser-side JavaScript, and drive a real browser.
For setup, install, or connection problems, read references/install.md.
Current ego lite (verified 0.4.7.4) preloads the legacy helpers below. page, browser, and taskSpaces are undefined. Do not import Playwright, invent facade names, or treat missing page.* as an install failure. If $HOME/.local/share/ego/ego-skills/SKILL.md disagrees with this file, follow that app-bundled skill for helper names and signatures, and keep this file's lifecycle, safety, and confirmation policies.
Run browser work with the Bash tool as ego-browser nodejs <<'EOF' ... EOF. Put the JavaScript directly in the heredoc; do not create a .js file first.
Inside a worktree-isolated session, the harness's worktree guard rejects heredocs as "too complex to verify that it stays inside the worktree." If that happens, write the script to a temp file and run ego-browser nodejs < /tmp/<name>.js instead.
A heredoc is only the JavaScript container. Default to one Bash invocation when the page will not navigate. Cap each heredoc at 60 seconds. After a navigation, stop and observe with snapshotText() or pageInfo() in a new heredoc. Start another command only for required user/external control, visual inspection that cannot happen in-process, or a process-level failure the script cannot recover from.