Calibrated Site SEO Audit
An SEO audit your agent can run on any public site and defend afterwards.
$49
Works with the AI Agents you already use
Calibrated Site SEO Audit
Example session with this skill installed
Audit https://www.python.org for a client pitch. Read a few pages, give me the scored table and an HTML report I can send them, and tell me what to fix first.
- Read your context and instructions
- Compiled the calibrated site seo
- Generated the document
Summary
Overall
67/100 across 19 scored areas, home page plus 3 sitemap pages read at one request per second. HTML report written to report.html, raw result to result.json. The site is technically clean; the losses are in structured data, social cards, crawlability and a heading-count problem.
$ node scripts/audit.mjs https://www.python.org --max-pages 3 --out report.html --json result.json
| Area | Score | Finding |
|---|---|---|
| Crawlability | 5/10 | robots.txt present, sitemap.xml MISSING |
| On-page basics | 6/10 | title "Welcome to Python.org" (21 ch); meta description 52 ch; canonical no; OpenGraph yes |
| Headings | 5/10 | 5 h1 elements, should be exactly one |
| Structured data | 3/10 | 1 JSON-LD block: WebSite only |
| Reviews / social proof | 0/10 | No review markup |
| Image alt text | 10/10 | 1/1 images have alt text |
| Analytics & pixels | 0/10 | No Google tag, no Meta pixel |
| Accessibility (quick pass) | 10/10 | No obvious issues |
| Favicon & manifest | 6/10 | favicon yes, web app manifest missing |
| Theme performance smells | 10/10 | No render-blocking or layout-shift smells |
| Social share cards | 6/10 | Missing twitter:card, twitter:image |
| Page weight & speed | 10/10 | HTML 51 KB, fetched in 205 ms |
| Local SEO | 0/10 | No LocalBusiness entity (ignore: not a local business) |
| E-E-A-T signals | 7/10 | Author byline, About page, entity schema present |
| AI crawler access | info | all 16 known AI crawlers allowed at root; llms.txt and ai.txt absent |
| Structured data placeholders | 10/10 | none |
| Content structure | 10/10 | no packed lists or over-long blocks |
| Text to HTML ratio | 10/10 | lowest page 12.6% |
| Titles and descriptions across pages | 10/10 | all unique |
| Page weight (measured) | 10/10 | first byte 108 ms, whole page 130 ms; 52 KB; 10 script tags, 1 third-party host |
Fix first (one-day items)
Sitemap: /sitemap.xml returns 404 while robots.txt is present. Publish one or point robots.txt at the real location. This also unlocks the sitemap freshness and link-health checks, which could not run.
2. Five h1 elements on the home page (<h1> is used for section headers). Demote four to h2.
3.
Meta description is 52 characters; the auditor's window is 70 to 165. Canonical link is absent; add <link rel="canonical">.
4. Twitter card: add twitter:card and twitter:image; OpenGraph is already present so the image exists.
Needs a developer
- Structured data is a bare
WebSiteblock; addOrganizationwith logo andsameAs, andBreadcrumbListon inner pages. - A web app manifest is optional but cheap.
Ignore
Local SEO and reviews: not a local business, no products.
Not verified
The home page is server-rendered so the content checks are valid. JavaScript-rendered sections were not executed. Core Web Vitals field data needs the PageSpeed API; the measured weight above is a lab proxy.
report.html
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
An SEO audit your agent can run on any public site and defend afterwards. One paced crawl (one request per second, so rate limits never show up as broken pages), then dependency-free checks over the HTML, scored 0 to 10 per area, with a Markdown table on stdout, a self-contained HTML client report and a JSON file for diffing runs.
What makes it different from an audit prompt: the thresholds were calibrated until the findings matched what a leading commercial site auditor flags on the same pages. "Paragraphs too long" is really three or more bullet glyphs packed into one paragraph; text-to-HTML fails under 10%; duplicate titles are compared after lowercasing and stripping punctuation; sitemap errors mean non-200, redirects, canonical elsewhere and noindex.
Areas covered: crawlability, on-page basics, headings, structured data and JSON-LD placeholders, reviews, image alt, analytics, accessibility basics, favicon and manifest, theme performance flags, social cards, hreflang, offer policies, Tag Manager container size, content structure, text-to-HTML ratio, duplicate titles, sitemap freshness, sampled link health, LocalBusiness schema, E-E-A-T signals, business identity consistency (name and phone across pages, optionally against the owner's values), measured home-page weight, and an informational AI-crawler read of robots.txt.
Also included: scripts/sitemap-check.mjs, a separate sitemap walker for sites with hundreds of pages that reports the "incorrect pages in sitemap" class plus title and description length and duplicates across every page fetched.
Node 20+, no packages, no API keys. Regex-based, no JavaScript rendering; the SKILL.md says how to report client-rendered sites honestly.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.
Regex parsing may fail on unusual markup. No JS rendering; SPA sites will score low. Core Web Vitals are lab proxies. Calibration is based on 2026 commercial auditor standards.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 19 days ago
- Software
Node.js 20 or later