seo-audit
SEO Audit
Overview
A page reaches search results only if a crawler can fetch it, it answers with HTTP 200, it carries content in a form the crawler reads, and no signal tells the engine to leave it out or prefer another URL. Most traffic losses come from one of those four breaking, usually after a redesign, a framework change or a configuration slip, and each one can be tested from a terminal.
This skill audits in that order: site-wide access, then a sample of pages, then rendering, speed and the source code that produced the fault. The output is a findings table in which every row has the command that showed the problem, so the user can rerun it after the fix. Rules quoted here follow Google Search Central as of October 2026; Bing and other engines read the same signals.
Instructions
1. Scope
Find out or ask: the canonical host (www or bare domain), what prompted the audit (a drop with a date, pages missing from the index, a pre-launch check), which page types earn the traffic, whether the repository is available, and whether the user can export data from Search Console. Do not sign in to anything: ask for the "Page indexing" export and the Performance report for the affected dates.
Work in a scratch directory and keep every output file; they are the evidence.