selenium-browser_skill

Overview

This skill starts a Selenium‑controlled Chrome (or Chromium) browser, opens a URL, takes a screenshot, and reports progress back to the caller. It supports headless mode and an optional proxy, and writes the screenshot file to a predictable location. Errors are captured and forwarded so callers know if something went wrong.

How this skill works

The skill runs a Python script that instantiates ChromeDriver with configurable options (headless, proxy, binary paths). It navigates to the provided URL, waits briefly for the page to load, captures a PNG screenshot, saves the image to disk, and prints a JSON payload with the status and file path. Any exceptions are printed as error text so the orchestrator can relay them.

When to use it

• Capture a visual snapshot of a web page for monitoring or archiving.
• Validate that a page renders correctly after a deploy or change.
• Collect screenshots for visual regression comparison or QA.
• Automate page capture from environments behind a proxy.
• Run headless browser tasks in CI or server environments.

Best practices

• Set CHROME_BIN and CHROMEDRIVER_PATH environment variables if binaries are not in standard locations.
• Prefer WebDriverWait instead of fixed sleep for reliable page load detection when dealing with dynamic pages.
• Run in headless mode for CI or headless servers to reduce resource usage.
• Use a dedicated output directory and unique filenames when taking multiple screenshots to avoid overwrites.
• Provide a proxy only when required and validate proxy accessibility beforehand.

Example use cases

• Scheduled job captures homepage screenshots every night for visual regression tracking.
• On-demand capture of a user's reported page to reproduce a rendering bug.
• CI step that opens a preview URL and archives a screenshot alongside build artifacts.
• Capture screenshots of pages that require routing through a corporate proxy for internal testing.