agentleFS
Sheldon-desouzaSheldon-desouza/claude-shopify-seo/skills/shopify-seo/SKILL.md

shopify-seo

Skill0 starsChanged 2 months agoMITView on GitHub

Write, check and publish Shopify product and blog copy that holds up for search, for AI assistants, and for the brand. Use for product descriptions, collection pages, blog posts, meta titles and descriptions, alt text, and Product schema. Triggers on: check my copy, optimise this description, is this SEO friendly, rewrite this product page, write a blog post for my shop, add schema, publish to Shopify.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 4 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Never invent social proof: No ratings, no review counts, no "loved by", no customer numbers, unless they exist in the brand profile with a …

    line 26
  2. Distinctive:

    Preserve the owner's voice: If a line already carries their phrasing, keep it.

    line 30
  3. Distinctive:

    Report honestly: If the copy is fine, say it is fine and change nothing.

    line 35
  4. Distinctive:

    Never return only a score.

    line 66

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Claude Shopify SEO

Turn a draft into copy that a shopper, a search engine and an AI assistant can all use, then publish it only when the owner has said yes to a change they can actually read.

The loop

  1. Load the brand profile from brand/<brand>.json. If none exists, run /shopify-seo brand first. Everything downstream is checked against this file. It is the only source of facts.
  2. Check the draft against reference/checks.md. Report pass, warn and fail per item.
  3. Fix what failed, changing as little as possible. Never rewrite a voice that already works.
  4. Re-check the fixed version and show what moved. If something still fails, say so plainly rather than declaring success.
  5. Show the owner a diff in plain English, not a JSON patch.
  6. Publish only on an explicit yes, or hand back text to paste.

Hard rules

  • Never invent a fact. Fabric, price, sizes, lead times, provenance and materials come from the brand profile or the live store. If a fact is missing, ask. Do not reason it out from the product name.
  • Never invent social proof. No ratings, no review counts, no "loved by", no customer numbers, unless they exist in the brand profile with a source.
  • Never claim a superlative that would need substantiating: best, finest, leading, number one. The brand profile's banned list is a floor, not a ceiling.
  • Preserve the owner's voice. If a line already carries their phrasing, keep it. You are editing, not auditioning.
  • A write to a live store is a mutation. It needs the diff shown and an explicit approval in the same conversation. Never publish on an implied yes, never batch-publish without per-item approval, never touch a product the owner did not name.
  • Report honestly. If the copy is fine, say it is fine and change nothing. A skill that always finds problems is a skill nobody trusts.

Routing

IntentRoute
Is this good, what is wrong with it/shopify-seo check
Make it better, optimise, rewrite/shopify-seo fix
Structured data, JSON-LD, rich results/shopify-seo schema
AI Overviews, ChatGPT, Perplexity, GEO, AEO, AI visibility/shopify-seo geo
Put it live, update the product, post it/shopify-seo publish
Set up a new brand or change the facts/shopify-seo brand

A bare /shopify-seo <text> runs check, then fix, then re-check, and stops for approval.

What "good" means here

Three audiences, and copy has to serve all three. reference/checks.md holds the detail.

  • The shopper. Can they tell what it is, what it costs, whether it fits, and when it arrives?
  • Search. Title and meta within length, one H1, alt text, internal links, the real search term used the way a person would say it.
  • AI assistants. Self-contained passages a model can lift and cite, specific facts rather than adjectives, and complete Product schema. Assistants read JSON-LD before they parse the page. Note that Google's own position is that this is still SEO rather than a separate discipline, so treat anyone selling GEO as a distinct service with caution. See /shopify-seo geo.

Output

Return a short report, then the improved copy in full, then the diff summary, then ask. Never return only a score. A number without the sentence that caused it is not useful to anybody.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.