OmO
Oh My OpenAgentv4.7.5

playwright

Browser automation skill backed by the Playwright MCP. Default provider for browser tasks, testing, and screenshots.

playwright is the default browser-automation provider. It auto-installs the Playwright MCP via npx and exposes the browser via MCP tools. Use it whenever the agent needs to click, fill, navigate, screenshot, or scrape a page.

Triggers

Browser tasks, testing, screenshots, "navigate to", "verify in browser".

Capabilities

  • Navigate, interact, fill forms, click elements.
  • Take screenshots and PDFs.
  • Wait for network requests.
  • Scrape content.

Switching to agent-browser

If you'd rather drive a CLI-based provider, swap the engine:

{ "browser_automation_engine": { "provider": "agent-browser" } }

agent-browser is documented separately under its own skill.

Source Notes

Aligned with upstream docs/reference/features.md#browser-automation-options. Disable via disabled_skills: ["playwright"].

On this page