agentleFS
adityash8adityash8/claude-seo-skills/.claude/skills/seo-ops/SKILL.md

seo-ops

Skill0 starsChanged 21 days agoMITView on GitHub

Use when you need an SEO execution workflow, research table, content brief, metadata variants, structured data, citation review, or bulk content plan.

6 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

10 rules. 0 are common, 10 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 6 edited versions we read keep these.

  • 10

    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:

    Use current official documentation for changing platform behavior, rich-result eligibility, crawler controls, and measurement definitions.

    line 15
  2. Distinctive:

    Every statistic needs a named source, date, and relevant context.

    line 16
  3. Distinctive:

    Use plain sentences, natural headings, and no em dashes in generated copy.

    line 19
  4. Distinctive:

    Read only the reference needed for the current job.

    line 24
  5. Distinctive:

    Do not present a citation, recommendation, referral, or conversion as interchangeable evidence.

    line 42
  6. Distinctive:

    Treat proposed metadata as drafts until the deployed HTML and canonical URL are checked.

    line 44
  7. Distinctive:

    Keep a baseline, changed-URL cohort, untouched comparison cohort where useful, and rollback copy for bulk work.

    line 45
  8. Distinctive:

    Never apply blind body-copy replacements to title or description fields.

    line 46
  9. Distinctive:

    Keep hostname and canonical distinctions during analysis.

    line 47
  10. Distinctive:

    Do not add a schema block, FAQ, statistic, product mention, or extra page merely to fill a template.

    line 59

“Common” is measured: kept by at least 6 in 10 of the 6 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

SEO Ops

Turn an SEO request into a finished, evidence-backed deliverable. This is an execution toolbox for operators with a defined site, audience, and goal.

Operating contract

  • Execute the requested operation when inputs and access are sufficient. Return a reusable prompt only when the user asks for one.
  • Confirm the target URL or keyword, audience, market, language, deliverable, and permitted actions. Ask only for missing inputs that affect the result.
  • Inspect the actual page and available search data. Distinguish observed findings from hypotheses and unavailable data.
  • Use current official documentation for changing platform behavior, rich-result eligibility, crawler controls, and measurement definitions. Record the verification date.
  • Every statistic needs a named source, date, and relevant context. Never invent a number, quote, testimonial, credential, or observation.
  • Mark unresolved evidence as [VERIFY: evidence needed] in working material. Resolve or remove it before calling a deliverable ready for publication.
  • Preserve uncertainty and source attribution. A vendor's claim is a claim, not independent proof.
  • Use plain sentences, natural headings, and no em dashes in generated copy.
  • Reading and drafting do not authorize publishing, changing settings, outreach, or bulk edits. Respect existing authorization and prepare a reviewable result before requesting any missing approval.

Choose the operation

Read only the reference needed for the current job.

JobReferenceRelated workflow
Search results, question demand, keyword clusters, content scopeResearch promptsseo-audit for a broader diagnosis
AI answer citations and search-surface prioritiesResearch promptsUse observed citations, not promised lift
Article, comparison, definition, or expert briefContent promptsverified-seo-article-writing for a sourced draft
Snippet rewrite, content refresh, or link planContent promptsinternal-linking for implementation detail
Answer summaries and citation-friendly rewritesCitation promptsSeparate citations from recommendations
Titles, descriptions, CTR review, schema, or image specificationsMetadata and schema promptsseo-quick-wins for site-wide prioritization
Full workflow, bulk processing, or pre-publication reviewWorkflow promptsKeep batches reversible and measurable

Evidence and measurement guardrails

  • A search result is a dated observation for a query, location, language, device, and session. Do not present one result as universal.
  • Label generated questions and keywords as hypotheses unless demand is observed in search results or supplied data.
  • Benchmark content by useful facts and answered questions. Do not turn competitor word counts, keyword density, or link counts into targets.
  • Separate valid schema vocabulary from a search platform's current display eligibility. Validate both before promising a rich result.
  • Do not present a citation, recommendation, referral, or conversion as interchangeable evidence.
  • Inspect snippet directives and crawler policies before changing them. Search crawling, training access, and user-triggered retrieval can have different controls. Verify the relevant product's current docs.
  • Treat proposed metadata as drafts until the deployed HTML and canonical URL are checked. A saved field does not prove what the public page serves.
  • Keep a baseline, changed-URL cohort, untouched comparison cohort where useful, and rollback copy for bulk work.
  • Never apply blind body-copy replacements to title or description fields. Preview structured fields separately and sample the rendered output.
  • Keep hostname and canonical distinctions during analysis. Aggregate rows only after verifying they represent the same page, and sum metrics instead of overwriting rows.
  • Falling impressions with stable position can have several causes. Inspect query mix, demand, tracking, and search features before deciding to rewrite.

Output contract

Return:

  1. Scope and evidence: inputs, sources, observation dates, limits, and assumptions.
  2. Finished artifact: the table, brief, copy, markup, or plan requested.
  3. Verification: checks completed, unresolved items, and any deployment boundary.
  4. Next action: one concrete recommendation, owner if known, success measure, and regression guardrail.

Do not add a schema block, FAQ, statistic, product mention, or extra page merely to fill a template. Keep only what serves the reader and is supported.

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.