AGENT SKILL

seo-audit

Use when the user wants to audit, review, or diagnose SEO issues on a site: rankings, technical SEO, on-page, backlinks, Core Web Vitals, site speed.

License 许可
MIT
最近更新
2026-11-14
Stars
★ 4
Forks
1

01 前言 Front Matter

name
seo-audit
version
1.0.0
triggers
[ "seo audit", "seo analysis", "technical seo", "site audit", "core web vitals", "seo issues", "site speed" ]
description
Use when the user wants to audit, review, or diagnose SEO issues on a site: rankings, technical SEO, on-page, backlinks, Core Web Vitals, site speed.

02 资源文件 2

03 SKILL.md 正文

When to Use

Use when the user wants to audit, review, or diagnose SEO issues on a site: rankings, technical SEO, on-page, backlinks, Core Web Vitals, site speed. Produces a priority-ordered fix list an agent or developer can execute from.

Workflow

1. Crawl Audit (Screaming Frog style)

Crawl the site and check every URL for:

  • Titles: present, unique, 50-60 chars, keyword-front-loaded. Flag missing, duplicate, or truncated.
  • Meta descriptions: present, unique, 150-160 chars (not a direct ranking factor, but drives CTR).
  • H1: exactly one per page, unique, contains the primary keyword. Flag multiple, missing, or duplicated H1s.
  • Canonicals: self-referencing where intended; flag canonicals pointing to 404s, redirects, or other domains.
  • robots.txt: reachable, no Disallow: / on important sections, not blocking CSS/JS.
  • sitemap.xml: valid XML, referenced in robots.txt, contains only 200-status URLs, no noindex pages.
  • Status codes: flag 404s (fix or 301), 5xx (server issues), 3xx chains (more than one hop), soft-404s (200 pages saying "not found").
  • Redirects: chains and loops; 301 (permanent) not 302 for moved pages; update internal links to final URLs. Commands: Screaming Frog CLI — java -Xmx4g -jar screamingfrogseospider.jar --crawl https://your-domain.com --output_folder ./crawl; or curl -sI https://your-domain.com/page per URL for status checks.

2. On-Page & Content

  1. Keyword mapping: map every important page to its target keyword (one primary + 2-3 secondary). Flag pages with no clear target and keywords with no page.
  2. Content gaps: for each keyword, compare your page against the top 5 ranking results — missing sections, thinner content, weaker internal linking. Prioritize gaps on money pages first.
  3. Internal linking: every page needs ≥2-3 internal links with descriptive anchor text; fix orphan pages (crawl "Inlinks = 0").
  4. Structured data: check JSON-LD (Product, Article, FAQ, Breadcrumb, Organization) via the Rich Results Test or curl -s https://your-domain.com/page | grep -c 'application/ld+json'; validate each schema against its spec.

3. Technical SEO

  • Core Web Vitals: LCP < 2.5s, INP < 200ms, CLS < 0.1 (75th percentile, mobile). Measure with the PageSpeed Insights API — https://www.googleapis.com/pagespeedonline/v5/runPagespeed?url=https://your-domain.com&strategy=mobile — or npx lighthouse https://your-domain.com --only-categories=performance.
  • Mobile: responsive viewport, tap targets, no horizontal scroll; verify in device mode.
  • HTTPS: valid cert (no expiry warnings), no mixed content (http resources on https pages).
  • Indexability: pages meant to rank are indexable (not noindexed, not blocked by robots); check via site:your-domain.com or Search Console URL Inspection.

4. Off-Page (Backlinks)

Pull the backlink profile (Ahrefs/Semrush/Moz, or Search Console Links report if no paid tool): referring domains, total links, anchor text distribution, link velocity. Flag toxic/spam domains (disavow only the clearly spammy ones), over-optimized anchor text, sudden link drops (penalty vs lost links — investigate), and top pages by referring domains (your linkable assets).

5. Priority-Ordered Fix List

Rank fixes by impact × effort:

  1. P0 (this week): indexability blockers (accidental noindex/robots blocks), crawl errors on money pages, broken checkout/critical 404s, mixed content.
  2. P1 (this month): title/meta/H1 issues, redirect chains, sitemap hygiene, LCP/INP/CLS fixes on the top 20 traffic pages.
  3. P2 (next quarter): content gaps, internal linking, structured data rollout, disavow toxic links. Deliver as a table: Issue | Pages affected | Evidence (URL/check) | Fix | Priority.

Pitfalls

  • Auditing only the homepage — crawl the whole site; half the issues hide on deep pages.
  • Confusing "crawlable" with "indexable" — a page can be crawled yet blocked from indexing (noindex/robots).
  • Treating lab data as user experience — use field data (CrUX) for the real picture.
  • "Fixing" 404s by deleting pages — 301 to the closest relevant page to preserve equity.
  • Over-disavowing backlinks — Google ignores most disavows and the wrong ones can hurt; only clearly spammy links.

Verification

  • Re-crawl after fixes: flagged counts for each check class drop to zero (or documented exceptions remain).
  • curl -sI confirms 301s land on final URLs with no chains; sitemap URLs all return 200 and are indexable.
  • PageSpeed/CrUX shows LCP < 2.5s, INP < 200ms, CLS < 0.1 on fixed pages (75th percentile, mobile).
  • Search Console: indexed pages trend up; every P0 item on the fix list is closed with evidence.

04 同仓库更多 Skills