Hubcap skill

Use hubcap directly from Claude Code or similar tools. The skill gives the agent knowledge of hubcap's 114 commands so it can automate Chrome on your behalf — navigating pages, filling forms, taking screenshots, running accessibility audits, and more.

Install

Add the marketplace and install the skill:

/plugin marketplace add tomyan/claude-skills
/plugin install hubcap@tomyan-skills

Or install the skill directly:

git clone https://github.com/tomyan/claude-skill-hubcap.git ~/.claude/plugins/hubcap

How it works

The skill keeps a small footprint in context — just the command categories and key patterns. When the agent needs detailed usage for a specific command, it runs hubcap help <command> to pull the full documentation on demand. This means you get accurate, complete help without burning context on 114 command docs.

What the agent can do

• Navigate to pages and wait for them to load
• Fill out forms and click buttons
• Take screenshots and export PDFs
• Scrape text, links, tables, and structured data
• Run accessibility audits
• Monitor network traffic and intercept requests
• Emulate mobile devices and set viewport sizes
• Execute JavaScript and evaluate expressions
• Assert page state for testing
• Record interactions for replay

Example

Ask the agent something like:

Take a screenshot of https://example.com on an iPhone 12

The agent will run:

hubcap setup launch
hubcap emulate iPhone-12
hubcap goto --wait https://example.com
hubcap screenshot --output example-iphone12.png

Prerequisites

Hubcap must be installed and Chrome must be available. The skill will use hubcap setup launch to start Chrome with the debug port open.

Source

tomyan/claude-skill-hubcap on GitHub.