Chrome Automation

Automates Chrome browser tasks by talking to your actual running Chrome session, not spinning up a headless instance. Built on agent-browser CLI. The key workflow is always checking existing tabs first before opening new ones, since the user's real browser already has their logins and session state. Handles the usual stuff like clicking, filling forms, taking screenshots, and extracting data. Can replay Chrome DevTools recordings, though you'll need to adapt selectors since IDs change between loads. Gets tricky with iframe-heavy sites like LinkedIn or Gmail because snapshots only see the main frame, so you'll sometimes need to drop down to eval or keyboard commands to reach embedded content.

skills.sh installs: 4K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add zc277584121/marketing-skills --skill chrome-automation --agent claude-code

Installs into .claude/skills of the current project.

Files

Skill: Chrome Automation (agent-browser)

Automate browser tasks in the user's real Chrome session via the agent-browser CLI.

Prerequisite: agent-browser must be installed and Chrome must have remote debugging enabled. See references/agent-browser-setup.md if unsure.

Core Principle: Reuse the User's Existing Chrome

This skill operates on a single Chrome process — the user's real browser. There is no session management, no separate profiles, no launching a fresh Playwright browser.