| name | technical-seo-audit |
| description | Run evidence-backed technical SEO audits for websites, docs sites, landing pages, and GitHub Pages properties. Use when Codex needs to inspect crawlability, title/meta/canonical tags, robots.txt, sitemap.xml, Open Graph/Twitter metadata, JSON-LD/schema, status codes, redirect/canonical issues, internal links, JavaScript rendering risk, Core Web Vitals boundaries, or technical SEO issues for a site. |
Technical SEO Audit
Audit crawlability and page-level technical SEO. Prefer deterministic evidence before recommendations.
Workflow
- Identify the target URL and scope: single page, docs site, sitemap URLs, or selected landing pages.
- Run the local page audit first:
python3 ../../scripts/site_meta_audit.py <url> --json
- For repository-backed sites, also run
github-repo-seo baseline collection.
- For multi-page or JavaScript-rendered sites, read
../../references/data-sources.md and choose Firecrawl, Playwright, or Lighthouse based on the proof needed.
- Classify every finding as
Confirmed, Likely, or Hypothesis.
Inspect
- HTTP status, redirects, canonical URL, and duplicate canonical risk.
<title>, meta description, robots meta, H1, headings, and indexability.
robots.txt, sitemap.xml, sitemap freshness, and discoverable URLs.
- Open Graph, Twitter cards, social preview assets.
- JSON-LD/schema presence and whether it matches page content.
- Internal links and orphaned page risk when sitemap/crawl data is available.
- Mobile/performance/CWV only when PageSpeed, CrUX, Lighthouse, or browser evidence is available.
Output
Use ../../references/report-template.md. Keep a concise action plan with owner-ready tasks or GitHub issues.
Rules
- Do not claim Core Web Vitals from HTML-only checks.
- Do not claim index coverage without Search Console or search-result evidence.
- Do not claim sitewide crawl health from a single page.
- Do not recommend hidden text, crawler-only copy, or model-only instructions.