agent-browser_skill

This skill helps you automate browser tasks for web testing, form filling, and data extraction with reliable element interactions.
  • Shell

206

GitHub Stars

2

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill fradser/dotclaude --skill agent-browser

  • SKILL.md12.2 KB
  • SYNC.md1.1 KB

Overview

This skill provides a concise command reference and usage patterns for browser automation tasks such as web testing, form filling, screenshots, and data extraction. It maps common flows into simple CLI commands and includes session, recording, and network controls to support repeatable developer and QA workflows.

How this skill works

You navigate to pages, take snapshots to obtain element references, and interact using those refs or semantic locators. The tool supports navigation, element interaction, information extraction, screenshots/PDF, video recording, network routing, and session management via straightforward CLI commands and global options. Use snapshots after major DOM changes to refresh refs and add --json for machine-readable output.

When to use it

  • Automating form submission, sign-in flows, and validation checks
  • End-to-end web testing with scripted user interactions and assertions
  • Capturing screenshots, PDFs, or video demos of web pages
  • Scraping or extracting structured data from dynamic pages
  • Debugging client-side issues with traces, console logs, and recordings

Best practices

  • Always run snapshot -i after navigation or visible DOM changes to get stable @refs
  • Use semantic locators (find role/text/label) for more resilient tests than brittle selectors
  • Persist authentication with state save/load to avoid repeated logins in flows
  • Prefer wait flags (--load, --url, --text) instead of fixed sleeps for stability
  • Run with --json in CI to parse outputs programmatically and assert results

Example use cases

  • Form automation: open page, snapshot -i, fill inputs by refs, click submit, wait for networkidle
  • Authenticated sessions: log in once, state save auth.json, reuse with state load in later runs
  • Visual capture: set viewport, open URL, screenshot --full or record start/stop for demos
  • Network testing: route requests to mock or abort third-party resources during tests
  • Parallel testing: run isolated browsers with --session to test multiple scenarios concurrently

FAQ

Use snapshot -i to produce @refs and re-run snapshot after navigation or DOM changes. Prefer semantic locators (find role/text/label/testid) when possible for resilience.

Can I run this in CI and parse outputs?

Yes. Add --json to commands like snapshot and get to receive machine-readable output that your CI can parse and assert against.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational