/capture

Capture a web page using sitecap — screenshot, accessibility tree, HTML, network, console, storage, and performance data.

Usage

/capture <url> [options]

Instructions

When this skill is invoked, run sitecap to capture the specified URL:

npx sitecap <url> --launch -o /tmp/sitecap-capture

After capture completes, read and summarize the key files:

1. Read the meta.json to confirm what was captured
2. Read screenshot.png to see the page visually
3. Read accessibility.txt for the page structure
4. Read console.json for any errors
5. Read performance.json for load metrics
6. Read network.json for request summary (count, failed requests)

Report findings concisely. If the user asked a question about the page, answer it using the captured data.

Options

Pass additional flags after the URL:

• --crawl --max-pages N — crawl the site
• -t screenshot,accessibility — capture specific types only
• -v 375x812 — mobile viewport
• --profile Default — use Chrome profile for authenticated pages

Examples

/capture https://example.com
/capture https://example.com -v 375x812
/capture https://myapp.com/dashboard --profile Default