SECURITY REVIEW
Not yet assessed
Review the original instructions and requested permissions before installing.
No security review is available for this catalog entry yet.
Web content discovery, URL crawling, and page inventory for accessibility audits. Use when scanning web pages, crawling sites for audit scope, or building page inventories for mul…
Web content discovery, URL crawling, and page inventory for accessibility audits. Use when scanning web pages, crawling sites for audit scope, or building page inventories for multi-page audits.
Review the original instructions and requested permissions before installing.
No security review is available for this catalog entry yet.
How clearly the skill guides your agent, how complete its workflow is, and how you can check the outcome.
No quality assessment is available for this catalog entry yet.
Original instructions from the publisher’s SKILL.md
# Web Scanning
## Supported Audit Methods
| Method | Tool | When to Use |
|--------|------|-------------|
| Runtime scan | axe-core CLI | Live URL available (dev server or production) |
| Code review | Agent specialists | Source code available in workspace |
| Both | axe-core + agents | Most comprehensive — catches issues from both angles |
## Runtime Scanning Commands
### axe-core CLI
```bash
# Single page scan (WCAG 2.2 AA)
npx @axe-core/cli <URL> --tags wcag2a,wcag2aa,wcag21a,wcag21aa
# Save results to JSON
npx @axe-core/cli <URL> --tags wcag2a,wcag2aa,wcag21a,wcag21aa --save ACCESSIBILITY-SCAN.json
# Multiple pages
npx @axe-core/cli <URL1> <URL2> <URL3> --tags wcag2a,wcag2aa,wcag21a,wcag21aa --save ACCESSIBILITY-SCAN.json
# With Chrome flags (headless)
npx @axe-core/cli <URL> --tags wcag2a,wcag2aa,wcag21a,wcag21aa --chrome-flags="--headless --no-sandbox"
```
### axe-core Tag Reference
| Tag | Standard |
|-----|----------|
| wcag2a | WCAG 2.0 Level A |
| wcag2aa | WCAG 2.0 Level AA |
| wcag21a | WCAG 2.1 Level A |
| wcag21aa | WCAG 2.1 Level AA |
| wcag22aa | WCAG 2.2 Level AA |
| best-practice | Best practices (not WCAG required) |
## Screenshot Capture
### capture-website-cli (preferred — lightweight)
```bash
# Full-page screenshot
npx capture-website-cli "<URL>" --output="screenshots/<page>.png" --full-page --type=png
# With viewport
npx capture-website-cli "<URL>" --output="screenshots/<page>.png" --full-page --width=1280 --height=720
# Mobile viewport
npx capture-website-cli "<URL>" --output="screenshots/<page>-mobile.png" --full-page --width=375 --height=812
# With delay for JS-rendered content
npx capture-website-cli "<URL>" --output="screenshots/<page>.png" --full-page --delay=3
```
### Playwright (fallback)
```bash
npx playwright screenshot --browser chromium --full-page --wait-for-timeout 3000 "<URL>" "screenshots/<page>.png"
```
## Page Discovery for Multi-Page Audits
### Crawl Depth Modes
| Mode | Behavior | Max Pages |
|------|----------|-----------|
| Current page only | Scan single URL | 1 |
| Key pages | User-provided list | User-defined |
| Full site crawl | Follow internal links | 50 (default cap) |
### Sitemap-based Discovery
```bash
# Check for sitemap
curl -s <BASE_URL>/sitemap.xml | head -20
# Extract URLs from sitemap
curl -s <BASE_URL>/sitemap.xml | grep -oP '<loc>\K[^<]+' | head -50
```
### Link-based Crawling
When crawling from a start URL:
- Follow only same-domain links
- Skip anchor links (#), mailto:, tel:, javascript: links
- Skip file downloads (.pdf, .zip, .doc, etc.)
- Respect robots.txt
- Cap at 50 pages unless user overrides
- Track visited URLs to avoid duplicates
## Framework Detection
Detect the framework from workspace files to enable framework-specific scanning:
| Indicator | Framework |
|-----------|-----------|
| `package.json` contains `react` | React |
| `package.json` contains `next` | Next.js |
| `package.json` contains `vue` | Vue |
| `package.json` contains `@angular/core` | Angular |
| `package.json` contains `svelte` | Svelte |
| `.jsx` / `.tsx` files present | React/Next.js |
| `.vue` files present | Vue |
| `angular.json` present | Angular |
| `.svelte` files present | Svelte |
| Only `.html` files | Vanilla HTML |
## Source Code File Patterns
When doing code review, scan these file patterns:
```bash
# HTML files
**/*.html
# Component files (framework-specific)
**/*.jsx
**/*.tsx
**/*.vue
**/*.svelte
**/*.component.ts
**/*.component.html
# Style files (for contrast/visual checks)
**/*.css
**/*.scss
**/*.less
**/*.module.css
# Config files (for framework detection)
package.json
next.config.*
nuxt.config.*
angular.json
svelte.config.*
tailwind.config.*
```
## Scan Configuration
### Web Scan Config File: `.a11y-web-config.json`
```json
{
"profile": "standard",
"wcagLevel": "AA",
"wcagVersion": "2.2",
"axeTags": ["wcag2a", "wcag2aa", "wcag21a", "wcag21aa"],
"maxPages": 50,
"screenshots": false,
"framework": "auto",
"ignore": {
"paths": ["node_modules/**", "dist/**", "build/**", ".next/**"],
"rules": []
}
}
```
### Scan Profiles
| Profile | Phases | Thoroughness |
|---------|--------|-------------|
| Quick | 1, 3, 4, 9 | Errors and critical only |
| Standard | All (1-9) | Errors and warnings |
| Deep | All + extras | All severities + animation, cognitive, touch |