# Automate Browser Workflows With WebDriver

Manual browser checks are slow and hard to repeat. This skill guides Claude, Codex, and Claude Code through WebDriver automation.

## Install

```bash
npx skillstore add emillindfors/browser-automation
```

## Metadata

- Status: approved
- Slug: emillindfors-browser-automation
- Version: 1.0.0
- Author version: 1.0.0
- Skillstore revision: r1
- Version status: valid
- Tree hash: 733ac5bb94ce4911c6cd658befbf610799348d5e40a25a017822457d6b270574
- Author: EmilLindfors
- GitHub username: EmilLindfors
- License: MIT
- Repository: https://github.com/EmilLindfors/rust-browser-mcp/tree/master/skill/
- Ref: 30c73eac2afe762f6aa9c4553158769369d47351
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: confirmation\_required
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: network, filesystem, env\_access, external\_commands
- Quality score: 38
- Quality tier: warning
- Public page: https://skillstore.pages.dev/skills/emillindfors-browser-automation
- Manifest: https://skillstore.pages.dev/api/skills/emillindfors-browser-automation/manifest

## Capabilities

- Navigate pages and control Chrome, Firefox, and Edge sessions through WebDriver.
- Find elements, click controls, enter text, and submit forms.
- Capture screenshots, page source, text, attributes, and current URLs.
- Run performance, memory, resource, and console log checks.
- Create and execute reusable automation recipes with parameters and retries.
- Start, stop, list, and health-check managed browser drivers.

## Use Cases

- Regression Test Web Flows: Automate login, form submission, navigation, and visual checks across supported browsers.
- Collect Structured Web Data: Extract page text, product lists, tables, and links from approved sites using repeatable selectors.
- Diagnose Page Performance: Measure load timing, memory usage, resource behavior, and console errors during realistic browser sessions.

## Prompt Templates

### Check a Page

```
Open [URL], wait for [selector], then report the page title, current URL, and visible status text.
```

### Test a Form

```
Use a test account to fill [form name] on [URL], submit it, and capture validation messages or success text.
```

### Extract Approved Data

```
From [approved URL], collect [fields] from each item, follow pagination safely, and stop after [limit] pages.
```

### Audit Performance Across Browsers

```
Run a multi-browser performance audit for [URL], compare load metrics, console errors, memory usage, and screenshot evidence.
```

## Limitations

- Requires a compatible WebDriver and local browser installation.
- Cannot bypass CAPTCHAs, access controls, or site terms.
- Browser actions may expose authenticated page data to the active AI session.
- Setup scripts may modify Claude Desktop configuration when run.

## Best Practices

- Use test accounts and approved sites for all automation.
- Wait for stable elements before clicking or extracting data.
- Store credentials outside recipes and clear sessions after authenticated work.

## Anti Patterns

- Do not automate sites without permission or against their terms.
- Do not save screenshots or page source containing sensitive data to shared paths.
- Do not rely on fixed sleeps when an element wait is available.

## Security Audit

- Audited at: 2026-07-05T14:41:38.132\+00:00
- Summary: Most static network and external-command matches are documentation examples, installer help text, or Markdown backticks and were dismissed. Real risks remain around broad Read/Bash/Write permissions, local screenshot and config writes, authenticated scraping, and browser-context JavaScript execution. No prompt injection attempt was found in the reviewed files.

## Stats

- Views: 186
- Downloads: 10
- Favorites: 0
- Popularity score: 0