Agent-Skills.md

Agent Skills: TabzChrome Browser Automation

Browser automation via 70 tabz MCP tools. Use when taking screenshots, filling forms, debugging network requests, testing responsive design, or using text-to-speech notifications.

UncategorizedID: GGPrompts/TabzChrome/browser

Repository

GGPrompts/TabzChrome
GGPromptsLicense: MIT
12010

Install this agent skill to your local

pnpm dlx add-skill https://github.com/GGPrompts/TabzChrome/tree/HEAD/plugins/tabz/skills/browser

Skill Files

Browse the full folder contents for browser.

Download Skill

Loading file tree…

plugins/tabz/skills/browser/SKILL.md

Skill Metadata

Name
browser
Description
"Browser automation via 70 tabz MCP tools. Use when taking screenshots, filling forms, debugging network requests, testing responsive design, or using text-to-speech notifications."

TabzChrome Browser Automation

Control Chrome via MCP tools for screenshots, interaction, debugging, and notifications.

Quick Start

Use mcp-cli to discover and call tools:

# Get tool schema (REQUIRED before calling)
mcp-cli info tabz/tabz_screenshot

# Call tool
mcp-cli call tabz/tabz_screenshot '{}'

Core Workflows

Screenshot a Page

# Get current tab info
mcp-cli call tabz/tabz_get_page_info '{}'

# Take screenshot
mcp-cli call tabz/tabz_screenshot '{}'
# Returns file path - use Read tool to view

Debug Network/API Issues

# 1. Enable capture BEFORE triggering action
mcp-cli call tabz/tabz_enable_network_capture '{}'

# 2. Trigger the action on page

# 3. Get failed requests (status >= 400)
mcp-cli call tabz/tabz_get_network_requests '{"statusMin": 400}'

# 4. Check console for JS errors
mcp-cli call tabz/tabz_get_console_logs '{"level": "error"}'

Test Responsive Design

# Emulate device
mcp-cli call tabz/tabz_emulate_device '{"device": "iPhone 14"}'

# Take screenshot
mcp-cli call tabz/tabz_screenshot '{}'

# Clear emulation
mcp-cli call tabz/tabz_emulate_clear '{}'

Fill and Submit Form

mcp-cli call tabz/tabz_fill '{"selector": "#email", "value": "test@example.com"}'
mcp-cli call tabz/tabz_click '{"selector": "button[type=submit]"}'

Notify User (TTS)

mcp-cli call tabz/tabz_speak '{"text": "Task complete"}'

Performance Profiling

mcp-cli call tabz/tabz_profile_performance '{}'
# Returns: DOM nodes, JS heap, event listeners, timing

DOM Tree Inspection

mcp-cli call tabz/tabz_get_dom_tree '{"maxDepth": 3}'

Tool Categories

| Category | Count | Key Tools | |----------|-------|-----------| | Screenshots | 2 | screenshot, screenshot_full | | Interaction | 4 | click, fill, get_element | | Network | 3 | enable_network_capture, get_network_requests | | DOM/Debug | 4 | get_dom_tree, get_console_logs, profile_performance | | Emulation | 6 | emulate_device, emulate_geolocation | | Audio/TTS | 3 | speak, list_voices, play_audio | | Tabs | 5 | list_tabs, open_url, switch_tab | | Cookies | 5 | cookies_get, cookies_list |

Important Notes

  • Always run mcp-cli info tabz/<tool> before calling
  • Use explicit tabId when possible - don't rely on "active" tab
  • Tab IDs are large integers (e.g., 1762561083)
  • tabz_screenshot cannot capture Chrome sidebar

References

See references/ for detailed workflows and full tool documentation:

Quick Guides

  • screenshot-workflows.md - Viewport vs full page screenshots
  • network-debugging.md - API request inspection
  • form-automation.md - Clicks, fills, selectors
  • tts-notifications.md - Audio feedback patterns

Full MCP Tool Documentation

  • core-tools.md - Tabs, screenshots, clicks, fills, DOM inspection
  • windows-groups.md - Windows, tab groups, multi-window workflows
  • network-downloads.md - Network capture, downloads, file operations
  • browser-data.md - History, sessions, cookies, bookmarks
  • profiles-and-plugins.md - Terminal profiles, Claude Code plugin management
  • advanced-tools.md - Emulation, performance profiling, notifications