Stacks Documentation
SkillsChangelog
TeamSponsorsPartnersPostcardwareAwesome StacksContributing

Browse

Copy page
Copy page as Markdown for LLMs
View as Markdown
View this page as plain text
Open in ChatGPT
Copy content and open ChatGPT
Open in Claude
Copy content and open Claude
Open in Perplexity
Search with Perplexity AI
On this page 6

stacks-browse · Engineering craft · model-invoked

Headless browser QA with nothing to install. It drives a Chromium-family browser already on the machine over the Chrome DevTools Protocol using only Bun, so navigation, screenshots, responsive checks, console and network monitoring and accessibility snapshots all work without Playwright or Puppeteer.

Component tests or browser QA

Reach for this skill when a check needs a real browser engine: real navigation, the CSS cascade, layout and element geometry, visibility, or screenshots. A virtual DOM cannot answer those, and an assertion about them written against one passes without checking anything.

For everything assertable through the DOM API — structure, events, forms, storage, web components — use component tests under bun test with very-happy-dom instead; see Testing. They run in-process and are fast enough for every save.

Note that this skill is a QA and diagnosis tool, not a test runner: there are no spec files, fixtures or reporters, and a scenario is expressed as CLI steps rather than as a committed test.

When to reach for it

  • Navigation
  • Screenshots
  • Responsive testing
  • console/network monitoring
  • Accessibility snapshots. Dependency-free
  • Driving a system browser over the Chrome DevTools Protocol using only Bun (no Playwright/Puppeteer)

Inside the skill

The sections an agent reads once the skill loads.

  • Browser discovery (no install step)
  • Default Stacks dev URLs
  • Commands
  • Stacks-Specific QA
  • Rules
  • Extending

Supporting files

Reference and scripts the skill reaches for on demand, rather than loading up front.

Using it

This one is model-invoked. Your agent reaches for it on its own when the task matches, and you can also call it by name:

/stacks-browse

Source: stacks-browse/SKILL.md. Shadow it for one project with app/Skills/stacks-browse/SKILL.md, then re-run buddy setup:ai. See Writing your own.