# Website Audit

> Lighthouse-style audit: perf, SEO, security, accessibility — driven by Playwright in the container.

- **Kind**: Shell tool
- **Category**: engineering
- **Owner**: Sentinel (/sentinel)
- **Default model**: sonnet
- **Cost class**: standard (Sonnet-class model. Default for most skills.)
- **Turn budget**: 8
- **Execution**: synchronous (result lands in the same turn)
- **Canonical URL**: https://app.51ultron.com/docs/skills/website-audit

## What it does

Runs ultron-render with Playwright against the target URL, capturing performance metrics (LCP, CLS, TTFB), scanning meta tags for SEO issues, checking common OWASP-Top-10 surface (CSP, HSTS, XSS, mixed content), and auditing accessibility (alt text, heading hierarchy, color contrast). Outputs a structured report with severity-ranked findings.

## When to use this

- user wants to audit their site or a competitor's
- user says 'audit this site' / 'how does X score on Lighthouse?' / 'check this site for security issues'
- user is shipping and wants pre-launch quality check

## When NOT to use this

- user only wants a screenshot → use ultron-screenshot directly
- user wants to fix issues, not just identify them → audit first, then minimal-patch

## Example prompts

- `audit stripe.com — perf, SEO, security`
- `how does my landing score on Lighthouse?`
- `run a full accessibility audit on this site`

## Inputs

- **url**: site to audit
- **focus**: optional comma-list: perf,seo,security,a11y (default: all)

## Output

Severity-ranked report (P0/P1/P2/P3) with metric values + concrete remediation per finding.

## Tools used

`shell_exec`

## Execution template

Shell tools dispatch through `shell_exec` with the following command template:

```bash
ultron-render {{url}} --audit={{focus}}; instrument browser-automation/SKILL.md flows for Lighthouse-style metrics
```

## Tags

`engineering`, `audit`, `verify`

## Keywords

audit, lighthouse, performance, seo, accessibility, owasp, security review