Skip to content

■ RECIPES // FIELD NOTE

Choose a Browser Automation Surface

Agent recipe for choosing Playwright, Puppeteer, browser-use, Playwright MCP, or no browser at all.

■ [!] ON THIS PAGE ▼

Use this recipe when the owner asks an assistant to use a website, verify UI, extract dynamic data, or debug a browser flow.

canonical sources

  1. Start with browser automation .
  2. Use Playwright MCP when the browser action belongs inside a coding-agent loop.
  3. Use browser agents for exploratory web task framing.
  4. Use sandboxing security when accounts, payments, sessions, or private data are involved.

agent task

Classify the browser task as stable, exploratory, protocol-heavy, coding-loop, API-first, or destructive. Recommend the smallest control surface and define the receipt before running anything.

required receipt

Return:

  • target URL;
  • chosen surface and why;
  • exact actions performed;
  • screenshot or accessibility snapshot;
  • extracted values;
  • final state;
  • uncertainty;
  • rollback or cleanup path;
  • canonical self.md URLs used.

refusal boundary

Do not complete purchases, send messages, delete data, submit irreversible forms, or change account settings without an explicit human approval gate.

Agent handoff

Use this comparison when the owner needs to choose Playwright, Puppeteer, browser-use, Playwright MCP, or no browser at all.

use when
  • the owner asks an assistant to use a website, verify UI, scrape a dynamic app, or debug a flow
  • the task needs screenshots, DOM state, traces, or a human approval gate
  • the owner needs to choose between Playwright, Puppeteer, browser-use, and Playwright MCP
receipt

screenshot or accessibility snapshot, URL, action log, extracted values, and final state

rollback

stop the session, clear the browser profile, undo the action, or require human sign-off

handoff prompt
Classify the browser task as stable, exploratory, protocol-heavy, coding-loop, API-first, or destructive. Recommend the smallest surface and define the receipt before running anything.

Open the task recipe

Agent for discovery, script for repeatability. See also llms.txt and agent context.