website-audit

name	website-audit
description	Audits a website for SEO, AI search performance, structured data, mobile usability, broken links, and social media cards. Fixes issues found. Use when the user mentions "audit website", "SEO", "fix search ranking", "AI search", "structured data", "social media cards", or "website performance".

Website Audit

Performs a comprehensive website audit and fixes issues affecting search visibility and AI discoverability.

Copy this checklist and track your progress:

Audit Progress:
- [ ] Step 1: Read guidelines
- [ ] Step 2: Update content correctness from codebase
- [ ] Step 3: Audit AI search readiness
- [ ] Step 4: Audit SEO and keywords
- [ ] Step 5: Audit crawling and indexing
- [ ] Step 6: Audit broken links and canonicalization
- [ ] Step 7: Audit mobile usability
- [ ] Step 8: Audit structured data
- [ ] Step 9: Audit social media cards
- [ ] Step 10: Audit For Unsubstantiated Claims
- [ ] Step 11: Audit Design Compliance
- [ ] Step 12: Test with Playwright
- [ ] Step 13: Report findings

• Read every document in each step carefully. Go deeper if needed.
• Check the outputted HTML/CSS/JavaScript AFTER the website is generated by the static content generator. - Don't just check the static content before the website is generated.
• Fix issues at the core where the static content templates are stored - not in the outputted HTML (e.g. _site)
• Never manually edit the generated website content directly

Step 1 — Read guidelines

Fetch and read each of these before auditing. These are the authoritative references for every step that follows.

• Google's guidance on using generative AI content
• Top ways to ensure content performs well in Google's AI experiences
• SEO Starter Guide

If the repo has a business plan or product-strategy doc, take it into account.

Identify the website source files in the repo. Determine the framework (static site generator, Next.js, Hugo, etc.) so you know where to find templates, metadata, and content.

Step 2 — Update content correctness from codebase

The website must accurately describe what the codebase actually does. Read the source code and cross-reference every claim on the website against the real implementation.

1. Inventory website claims — Identify every feature, capability, diagnostic code, supported construct, and technical detail mentioned on the website.
2. Read the codebase — Examine the relevant source files (checker rules, LSP handlers, diagnostic definitions, supported features) to determine what the code actually implements today.
3. Fix inaccuracies — For each claim on the website:
   • If the feature exists in the codebase, ensure the description matches the actual behavior.
   • If the feature does not exist yet, remove the claim or clearly mark it as planned/upcoming.
   • If the codebase supports something the website doesn't mention, consider adding it.
4. Verify diagnostic codes — Ensure any diagnostic codes (BSK-E#### / BSK-W####) listed on the website match the codes defined in the codebase.
5. Check version/compatibility claims — Confirm Python version targets, supported platforms, and dependency information match Cargo.toml, CI config, and docs.

Fix issues directly in the website source files. Do not overstate capabilities — accuracy builds trust with users and search engines.

Step 3 — Audit AI search readiness

Apply the guidance from the AI search article. Check:

1. Content quality — Is content original, expert-level, and comprehensive? Flag thin or duplicated pages.
2. Clear structure — Do pages use descriptive headings, lists, and concise answers to likely questions?
3. Entity clarity — Are key terms, products, and concepts defined clearly so AI can extract them?
4. Freshness signals — Are dates, update timestamps, and authorship present?

Fix issues directly in the source files. For each fix, note what changed and why.

Step 4 — Audit SEO and keywords

1. Search Google Trends for trending keywords related to the website's content.
2. Review each page's <title>, <meta name="description">, and <h1> tags.
3. Check for keyword opportunities — can trending terms be naturally inserted into headings, descriptions, or body content?
4. Verify each page has a unique, descriptive title (50-60 chars) and meta description (150-160 chars).
5. Check image alt attributes describe the image content and include relevant keywords where natural.

Apply the SEO Starter Guide principles. Fix issues directly.

Step 5 — Audit crawling and indexing

Reference: Overview of crawling and indexing topics

1. robots.txt — Locate and review it. Verify it doesn't block important pages. Reference: robots.txt spec
2. Sitemap — Locate the sitemap (or sitemap index). Verify all important pages are listed and no dead URLs are included. Reference: Sitemap guidelines
3. Meta robots tags — Check for unintended noindex or nofollow directives on pages that should be indexed.

Note: robots.txt and sitemaps are often auto-generated. If so, check the generator config rather than the output file.

Step 6 — Audit broken links and canonicalization

Reference: What is canonicalization

1. Check all internal links resolve to valid pages (no 404s).
2. Verify <link rel="canonical"> tags are present and point to the correct URL.
3. Check for duplicate content accessible via multiple URLs (with/without trailing slash, www vs non-www).
4. Verify redirects use 301 (permanent) not 302 (temporary) where appropriate.

Step 7 — Audit mobile usability

Reference: Mobile-first indexing best practices

1. Verify the <meta name="viewport"> tag is present and correctly configured.
2. Check that content is identical between mobile and desktop (mobile-first indexing requires this).
3. Verify touch targets are adequately sized (min 48x48px).
4. Check font sizes are readable without zooming (min 16px body text).

Step 8 — Audit structured data

Reference: Structured data guidelines

1. Check for existing JSON-LD <script type="application/ld+json"> blocks.
2. Verify the structured data matches the page content (no misleading markup).
3. Add missing structured data where appropriate:
   • Organization/Person on the homepage
   • Article/BlogPosting on blog posts (with author, datePublished, dateModified)
   • BreadcrumbList for navigation
   • FAQ for pages with question/answer content
4. Validate JSON-LD syntax is correct.

Step 9 — Audit social media cards

Reference: Implementing Social Media Preview Cards

Check every page template includes:

Open Graph (Facebook/LinkedIn):

• og:title, og:description, og:image, og:url, og:type

Twitter Card:

• twitter:card, twitter:title, twitter:description, twitter:image

Verify og:image dimensions are at least 1200x630px. Fix missing or incomplete tags.

Step 10 - Audit For Unsubstantiated Claims

Ensure that all claims are backed up with a link to a reputable source. As an example, this claim isn't valid as content unless it links to an authority that found this through research

Research shows teams with strong DevEx perform 4-5x better across speed, quality, and engagement

Search for the authoritative URL and add a link to the URL. If it is not available, change the claim to something that can be substatiated.

Step 11 — Audit Design Compliance

If the repo has a design system (design tokens, component docs, or design screens), review it and check the site for compliance. Otherwise, audit the live site's visual consistency directly (spacing, type scale, color usage, component reuse).

Step 12 — Test with Playwright

Build and run the website locally using make website-run (or the project's equivalent dev server command).

Desktop tests (1280x720):

1. Navigate to the homepage — take a screenshot.
2. Navigate to each major section — verify pages load without errors.
3. Check the browser console for JavaScript errors.
4. Verify all navigation links work.

Mobile tests (375x667, iPhone SE):

1. Resize the browser to mobile dimensions.
2. Navigate to the homepage — take a screenshot.
3. Verify the layout is responsive (no horizontal overflow, readable text).
4. Test navigation menu (hamburger menu if applicable).

If any page fails to load or has console errors, fix the issue and retest.

Step 13 — Report findings

Summarize the audit results:

## Website Audit Report

### Fixed
- [List each issue fixed with file and line reference]

### Warnings (manual review needed)
- [Issues that need human judgment]

### Passed
- [Areas that passed audit with no issues]

### Screenshots
- [Reference Playwright screenshots taken]

Rules

• Fix issues directly — don't just report them. Only flag issues as warnings when they require human judgment (e.g., content tone, keyword selection).
• One step at a time — complete each step before moving to the next.
• Preserve existing content — improve structure and metadata without rewriting the author's voice.
• No keyword stuffing — keywords must read naturally in context.
• Respect the framework — edit templates/configs, not generated output files.