agent-browser

Automates browser interactions for testing, verification, and screenshot capture via CLI.

1|Updated Apr 25, 2026
One-click install
npx skills add https://github.com/rczamor/rz-emdash --skill agent-browser-rczamor
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/rczamor/rz-emdash/tree/main/skills/agent-browser
Command: npx skills add https://github.com/rczamor/rz-emdash --skill agent-browser-rczamor

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Verifying frontend changes and testing web UI interactions manually is slow and error-prone. This Skill lets an AI agent drive a real browser to open pages, inspect the accessibility tree, fill forms, click elements, and capture screenshots for verification and debugging. ## Core Features & Use Cases - Accessibility-Tree Snapshots: Get interactive element refs (@e1, @e2) to precisely target clicks, fills, and selections without brittle selectors. - Form & Flow Testing: Fill inputs, select dropdowns, check checkboxes, and click buttons to verify complete user flows end to end. - Screenshots & Debugging: Capture viewport or full-page screenshots, view console logs and page errors, highlight elements, and run JavaScript with eval. - Use Case: After modifying a frontend page, open the local dev server, snapshot the interactive elements, fill and submit a form, then screenshot the result to confirm the change renders correctly. ## Quick Start Open http://localhost:4321 in the browser, take an interactive snapshot, and verify the page renders correctly with a screenshot.

Frequently Asked Questions about agent-browser

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate browser testing of web UI changes?▼

Open the target URL with agent-browser, take an interactive snapshot to get element refs, then use click, fill, and select commands against those refs. Finish with a screenshot to visually verify the result.

How to fill and submit a form using browser automation?▼

Take a snapshot with the -i flag to list interactive elements, then use fill with each field's @ref to enter values and click the submit button's ref. Use wait after actions that trigger page loading.

Can browser sessions persist login state between commands?▼

Yes, named sessions keep cookies and storage between commands. Pass --session with a name to each command or set the AGENT_BROWSER_SESSION environment variable to reuse authenticated state.

Why can't I load a large screenshot file?▼

Screenshot files larger than 2MB cannot be loaded by agent tools. Check the image size first and reduce it with a tool like sips or ImageMagick before attempting to load it.

How do I debug when browser automation is not working?▼

Run with the --headed flag to watch the browser window, then use console and errors commands to view page logs and errors. You can also highlight an element by ref or run JavaScript with eval to inspect state.