Doesn't Lighthouse or PageSpeed Insights already cover this?

They cover some of it per page, in a browser. They don't check robots.txt for Googlebot-Image, noindex headers, extension-vs-content-type, image-sitemap tags or alt-text quality across a whole sitemap, and they don't fail a CI job. Keep using them for Core Web Vitals — this is the image-specific, sitewide, scriptable layer.

Will it write my alt text?

It suggests alt text only when it can see the image or you describe it, and says so otherwise. It never derives alt text from a filename.

Does it convert images to WebP or AVIF?

No. It flags large legacy-format files as informational and tells you what to change; the conversion happens in your own toolchain.

What about JavaScript-rendered images?

The checker reads the raw HTML. Template placeholders and JavaScript-only galleries are reported as 'verify in the rendered DOM', not guessed.

Will you keep it updated?

Updates for 12 months from purchase, through your Gumroad library and logged in a changelog — including when Google changes its requirements. After that the version you have keeps working; further updates are not included.

Products → Image SEO & Optimization Skill Pack for Claude Code

Image SEO & Optimization Skill Pack for Claude Code

A Claude Code skill pack that audits the images on your site the way Google Images and the browser see them — alt text, filenames, formats, srcset, missing dimensions, lazy-loading mistakes, og:image, image sitemaps, and (with live checks) noindex headers, robots.txt blocks and heavy files — then gives the exact HTML change for each finding. Plus a pass/fail guard for CI. Zero dependencies, €49 once.

Get it on Gumroad — €49

v1.0.0 — live. Tested on real live sites (this one, allbirds.com, web.dev, bbc.com, Wikipedia) plus labelled broken/fixed fixtures and a local server that exercises every live check. On this site it found that no page had an og:image — fixed the same day. The first version also gave three false alarms on those real sites, all fixed before release and listed in the testing record. The worked outputs ship in the download (esempio/).

The problem this solves

Image SEO fails quietly. Nothing errors; the pictures just never show up in Google Images, or they arrive late and shove the page around. An image with only data-src, a <picture> with no <img> fallback, or a file served with X-Robots-Tag: noindex is invisible to Google — and your site looks fine in the browser. Alt text, which Google's docs call the most important image metadata, is missing, or "image", or a list of keywords. No width/height means layout shift; a lazy-loaded hero image arrives late; a folder blocked for Googlebot-Image in robots.txt hides everything in it. Each is a one-line fix — if you know it's there.

What's inside

Skill What it does

image-seo-audit Runs over a page, a whole sitemap or your local build folder and reports Blocking / Should-fix / Info: no src, unsupported formats, missing / generic / keyword-stuffed alt, image-only links with no anchor text, generic filenames, missing dimensions, srcset without sizes, lazy-loaded hero, CSS-background images (not indexed), missing or relative og:image, deprecated image-sitemap tags. With --check-urls it also fetches every image: status, content-type vs extension, weight against your budget, X-Robots-Tag, and robots.txt for Googlebot-Image. Claude then gives the exact change per finding and reads a sample of your alt texts for what a script can't judge.

image-seo-guard The same checks as a binary pass/fail gate: --assert exits non-zero on a regression. Run it on your build folder before every deploy.

Skill	What it does
`image-seo-audit`	Runs over a page, a whole sitemap or your local build folder and reports Blocking / Should-fix / Info: no `src`, unsupported formats, missing / generic / keyword-stuffed alt, image-only links with no anchor text, generic filenames, missing dimensions, `srcset` without `sizes`, lazy-loaded hero, CSS-background images (not indexed), missing or relative og:image, deprecated image-sitemap tags. With `--check-urls` it also fetches every image: status, content-type vs extension, weight against your budget, `X-Robots-Tag`, and robots.txt for Googlebot-Image. Claude then gives the exact change per finding and reads a sample of your alt texts for what a script can't judge.
`image-seo-guard`	The same checks as a binary pass/fail gate: `--assert` exits non-zero on a regression. Run it on your build folder before every deploy.

Plus one dependency-free helper script, image-seo-check.mjs (Node 18+), an install guide, and worked examples against real live sites.

What it is not

It doesn't compress or convert your images. It tells you which ones to fix and how.
It checks mechanics, not meaning: it can tell that alt text is missing or generic, not whether it accurately describes the photo. It will not invent alt text for an image it hasn't seen.
It reads the raw HTML. Images added by JavaScript are flagged as "check the rendered page", not guessed at.

Why €49

If your time is worth €50 an hour, the pack pays for itself the first time it saves you an hour of crawling your own templates for missing alt text and dimensions — or the first time the guard stops a deploy that would have shipped a lazy-loaded hero image or an upload blocked by robots.txt.

What you get

2 skills and 1 dependency-free helper script
Install guide (~2 minutes) and a testing record with the Google doc sources behind every rule
Worked examples on real live sites (mulino.tools, allbirds.com, web.dev, bbc.com, Wikipedia), a labelled synthetic broken/fixed fixture set and a local server that exercises every live check
12 months of updates — a bounded, honest commitment, not "lifetime"
30-day refund, no questions asked, through Gumroad

Requirements

Claude Code (free)
Node.js 18+. No npm dependencies, no API keys, no account.
Pages reachable over HTTP(S), or a local build folder of HTML files.

FAQ

Doesn't Lighthouse or PageSpeed Insights already cover this?: They cover some of it per page, in a browser. They don't check robots.txt for Googlebot-Image, noindex headers, extension-vs-content-type, image-sitemap tags or alt-text quality across a whole sitemap, and they don't fail a CI job. Keep using them for Core Web Vitals — this is the image-specific, sitewide, scriptable layer.
Will it write my alt text?: It suggests alt text only when it can see the image or you describe it, and says so otherwise. It never derives alt text from a filename.
Does it convert images to WebP or AVIF?: No. It flags large legacy-format files as informational and tells you what to change; the conversion happens in your own toolchain.
What about JavaScript-rendered images?: The checker reads the raw HTML. Template placeholders and JavaScript-only galleries are reported as 'verify in the rendered DOM', not guessed.
Will you keep it updated?: Updates for 12 months from purchase, through your Gumroad library and logged in a changelog — including when Google changes its requirements. After that the version you have keeps working; further updates are not included.

AI disclosure

The skills are written instructions for an AI coding assistant and were authored with AI assistance. The helper script is conventional code and reports only what it finds in your pages — no invented metrics, rankings or traffic claims anywhere in the product or on this page.

Get it on Gumroad — €49

Eighth in the mulino.tools technical-SEO set for Claude Code: the Technical SEO Skill Pack audits and fixes an existing site, the Site Migration & Redirects Skill Pack covers the replatform/relaunch moment, the Rendering & JavaScript SEO Skill Pack covers the raw-vs-rendered gap JS frameworks introduce, the Robots, Sitemap & Crawl-Directives Skill Pack covers robots.txt vs sitemap.xml, the Hreflang & International SEO Skill Pack covers hreflang reciprocity, the Core Web Vitals & Performance Budget Skill Pack measures real LCP/CLS, the Rich Results & Structured Data Skill Pack checks your structured data per rich-result feature, and this one audits the images themselves plus a CI guard.