Intelligence · Claude Code skill

SEO/GEO Audit skill and the Yalc Framework

The client-facing SEO audit deliverable. Comparative mode is the killer feature: re-audits a quarter later show score deltas, resolved findings, and new gaps overlaid on the prior report.

Yalc Fit Score
9/10
License
MIT (Yalc)
Output
Vercel report
Mode
New + comparative
Last reviewed
2026-04-29
Trigger phrases

Say this to fire the SEO/GEO Audit skill

Any of these natural language phrases activates the skill inside Claude Code.

run SEO audit for <client>
audit <client> website
SEO report for <url>
GEO audit for <client>
seo audit
client audit
re-audit
follow-up audit
What it does

SEO/GEO Audit, plainly

The SEO/GEO Audit skill runs a full audit on a client website (technical SEO, content, schema, GEO/AI search readiness, sitemap, etc.), then deploys an interactive Vercel report under a custom subdomain (e.g. `audit-clientname.vercel.app`). Output is structured JSON consumed by the report template at `~/Desktop/earleads-seo-report/`.

Comparative mode is unique to this skill: when a prior audit JSON exists for the same client, the skill auto-archives the old audit and produces a `comparative` block (score deltas, resolved findings, still-open findings, new findings). The new audit deploys to a `-v2` or `-v3` subdomain so prior audits stay live for client reference.

Where it slots in

Position in the GTM operating system

Intake
→
Enrich
→
Score
→
Route
→
Draft
→
Send
→
Listen

The SEO Audit skill sits at the **listen** node when the audit is the deliverable. It also produces actionable recommendations that flow into downstream content or dev work, so it's adjacent to **draft**.

The skill is most powerful when run quarterly. Comparative mode shows clients what improved (justifies the retainer), what's still open (sets the next quarter's plan), and what's newly broken (surfaces unexpected regressions).

The Yalc Framework

Running the SEO/GEO Audit skill end to end

Workflow position

The audit deliverable. Yalc runs this on demand at the start of each engagement (baseline audit) and quarterly thereafter (comparative audits).

Required inputs

  • → Client website URL
  • → [object Object]
  • → [object Object]

Outputs

  • → Structured JSON for the Vercel report template
  • → Deployed Vercel preview URL
  • → Comparative block when a prior audit exists (delta scores, findings diff)
  • → [object Object]

Chaining recommendations

UpstreamYalc prompt with client URL → earleads-seo-audit
DownstreamAudit findings → content-strategist (content gaps) or dev work (technical fixes)

Anti patterns to avoid

Don't run the audit without enriching context files. Generic audits read as auto-generated. Battlecards plus keyword files plus prior work files dramatically improve output quality.
Don't deploy the audit to production before reviewing locally first. Yalc previews on Vercel; client-facing prod deploys happen only after the AM has read end-to-end.
Don't overwrite prior audits. Comparative mode auto-archives by date and deploys to a versioned subdomain so prior audits stay live.
Operator take

Pros, cons, who it's for

Pros

  • Client-facing deliverable lives at a real URL, not in a doc
  • Comparative mode shows progress quarter over quarter
  • Auto-archival of prior audits prevents accidental overwrite
  • Customizable with per-client context files (battlecards, keywords)
  • Vercel deploy makes the report shareable and link-stable

Cons

  • Vercel deploy requires VERCEL_TOKEN setup per Earleads workspace
  • First audit per client is slow (full crawl + analysis takes 45 to 90 minutes)
  • Comparative mode requires the prior audit JSON to exist; first audits are baseline-only
  • Report template is HTML; rich edits require dev work, not just content updates

Who it's for

  • Earleads SEO/GEO clients (active retainer engagements)
  • Agencies running quarterly SEO audits as part of retainer scope
  • Operators wanting a credible client-facing SEO deliverable without ahrefs-style raw data dumps
Dependencies

What this skill expects to find

Other skills

MCP servers

Environment variables

Report template repo at `~/Desktop/earleads-seo-report/`. The skill deploys via the Vercel CLI. Prior audits archive as `audit-results-YYYY-MM-DD.json` automatically. v2.1.0 introduced comparative mode (April 24, 2026) with the DataScaleHR baseline audit going from score 37 to 65.

Related

The SEO/GEO Audit ecosystem inside Yalc

Web Browsing skill
Provides the underlying Firecrawl wrapper used by this skill for site crawling.
→
Alternatives

Skills that overlap

Ahrefs or Semrush exports
Switch when the deliverable is raw data, not interpreted findings with action items.
→
Manual SEO consultant audit
Switch when stakes justify a senior SEO consultant's bespoke deliverable.
→
Lumar or Sitebulb crawl
Switch when you specifically need enterprise-grade crawl features.
→
FAQ

Frequently asked

How long does an audit take?

First audit is 45 to 90 minutes depending on site size. Comparative re-audits are faster (30 to 60 minutes) because the framework is already known and the diff focus narrows the analysis.

What does comparative mode produce?

A `comparative` block in the JSON output with score deltas (overall and per-section), resolved findings (gone since last audit), still-open findings (still present), and new findings (introduced since last audit). The Vercel template renders this as a progress overlay.

Where does the audit deploy?

A custom subdomain like `audit-clientname.vercel.app` for production. Comparative re-audits deploy to `audit-clientname-v2.vercel.app` (and v3, v4, etc.) so prior versions stay live.

Can the audit be customized per client?

Yes. Drop battlecards, keyword files, content strategy docs, and prior SEO work into the client's `01_Projects/Clients/Active/<client>/` folder. The skill reads them at runtime and grounds findings in the client's specific context.

What's the first DataScaleHR audit example?

First audit was March 18, 2026, score 37. Re-audit April 24, 2026 score 65 (+28). Comparative overlay shows AI Search +59, Schema +37, Images +36, Content +32. 13 findings resolved, 20 still open. Live at `audit-datascalehr-v2.vercel.app`.

Does it cover GEO (AI search) explicitly?

Yes. The audit checks AI crawler accessibility (GPTBot, ClaudeBot, PerplexityBot), llms.txt compliance, brand mention signals, and passage-level citability. GEO is a first-class section, not bolted on.

Get the SEO/GEO Audit skill

Clone the Yalc skill set, drop in your env, run from your next Claude Code session.

gh repo clone Othmane-Khadri/YALC-the-GTM-operating-system && cp -r YALC-the-GTM-operating-system/.claude/skills/earleads-seo-audit ./.claude/skills/
View on GitHub We deploy it for you
Last reviewed 2026-04-29 by Othmane Khadri