audit-website_skill

Overview

This skill audits websites for SEO, performance, security, accessibility, content, and technical issues using the squirrelscan CLI. It runs a browser-like crawl against your site and returns LLM-optimized reports that include a 0–100 health score, category breakdowns, and prioritized, actionable recommendations. Use it to discover broken links, meta tag problems, rendering issues, security gaps, and other maintainability items across 230+ rules.

How this skill works

The skill invokes the squirrel CLI to crawl and analyze pages according to a chosen coverage mode (quick, surface, full). It evaluates pages across 21 rule categories and assigns severity levels and ranks, then generates a compact LLM-format report with overall score, per-category breakdowns, affected URLs, and repair suggestions. Results are cached in a local project database and support diffs between audits for regression tracking.

When to use it

• Run an initial health check before a launch or migration
• Debug technical SEO issues like canonicalization, redirects or meta tags
• Locate and prioritize broken internal and external links
• Validate structured data, Open Graph, and Twitter card markup
• Monitor regressions after code or content changes using diff mode

Best practices

• Prefer auditing the live site for true rendering and performance data
• Start with a surface scan, then run a full scan for deep analysis
• Use the llm output format for concise, actionable reports
• Group repeatable content fixes (alt text, headings) and parallelize them
• Re-audit after each fix batch and compare before/after scores

Example use cases

• Generate a baseline audit and score for an e-commerce site before a marketing campaign
• Run daily quick audits in CI to catch regressions in the build pipeline
• Find and fix accessibility issues (alt text, contrast, keyboard focus)
• Identify leaked secrets, mixed content, and missing security headers
• Produce an LLM-friendly export to feed into downstream automation or task lists