agent-browser | Skill Performance & Reviews | TopRankSkills

TopRank Skills

Home / Skills / tools / agent-browser

agent-browser

maintained by wormholecowboy

star 1 account_tree 0 verified_user MIT License
bolt View GitHub

name: agent-browser description: Headless browser automation via CLI for web scraping, UI testing, and automated workflows. Use when tasks require navigating websites, interacting with page elements (clicking, typing, form filling), taking screenshots, verifying page content, or performing end-to-end browser automation. Assumes agent-browser is installed globally.

Agent Browser

Control a headless browser via CLI. Use --json flag for machine-readable output.

Core Workflow

  1. Navigate to target URL
  2. Snapshot to get accessibility tree with element refs (@e1, @e2, etc.)
  3. Act using refs for deterministic element selection
  4. Re-snapshot after page changes
agent-browser open "https://example.com"
agent-browser snapshot -i -c    # interactive elements only, compact
agent-browser click @e3
agent-browser snapshot -i -c    # re-snapshot after interaction

Selectors

Prefer refs from snapshots for reliability:

  • @e1 - Ref from accessibility tree (most reliable)
  • text=Submit - By text content
  • #id / .class - CSS selectors
  • xpath=//button - XPath

Common Patterns

Form Interaction

agent-browser fill @e5 "user@example.com"
agent-browser fill @e6 "password123"
agent-browser click @e7
agent-browser wait --url "/dashboard"

UI Verification

agent-browser is visible "#success-message"
agent-browser get text ".error-message"
agent-browser get count ".list-item"
agent-browser get attr @e2 "disabled"

Screenshots & PDFs

agent-browser screenshot ./page.png
agent-browser screenshot --full ./fullpage.png
agent-browser pdf ./document.pdf

Wait Patterns

agent-browser wait "#loading"              # Wait for element
agent-browser wait 2000                    # Wait ms
agent-browser wait --text "Success"        # Wait for text
agent-browser wait --load networkidle     # Wait for network

Sessions (Parallel Isolation)

agent-browser --session test1 open "https://example.com"
agent-browser --session test2 open "https://other.com"
agent-browser session list

Authentication

# Set headers for API auth
agent-browser --headers '{"Authorization": "Bearer token"}' open "https://api.example.com"

# Save/restore auth state
agent-browser state save ./auth.json
agent-browser state load ./auth.json

Network Mocking

agent-browser network route "**/api/users" --body '{"users": []}'
agent-browser network route "**/ads/*" --abort

Snapshot Tips

Flag Purpose
-i Interactive elements only (buttons, inputs, links)
-c Compact output, removes empty containers
-d 3 Limit depth to 3 levels
-s "#main" Scope to specific container

Always use -i -c for cleaner output when interacting.

Reference

Full command reference: references/commands.md

chat Comments (0)

chat_bubble_outline

No comments yet. Be the first to share your thoughts!

Skill Details

GitHub Stars 1
GitHub Forks 0
Created Jan 2026
Last Updated 5 months ago
tools tools productivity tools

Related Skills

planning-with-files
chevron_right
agent-browser
chevron_right
building-agents
chevron_right
notebooklm
chevron_right
grafana
chevron_right

Build your own?

Join 12,000+ developers contributing to the Claude ecosystem.