`/local-seo-audit`

Backed by `.claude/context/ai-seo/playbook/01-local-seo-audit.md` (+ `00-field-notes.md`).

Placeholders like ACME Agency, <id> and you@example.com mark values that are per-agency — your install fills them with YOUR clients and accounts. If a section references a helper script you don't have yet, it ships with that workflow's install.

/local-seo-audit — AI-Powered Local SEO Audit

Backed by .claude/context/ai-seo/playbook/01-local-seo-audit.md (+ 00-field-notes.md). Method source: Caleb Ulku, "AI SEO Mastery Pro". Shared I/O runner: ACME Agency/scripts/seo_run.mjs.

When to trigger

"Run a local SEO audit for ACME Agency", "audit the SEO for this business", "/local-seo-audit ACME Agency"
A prospect/client asks "why aren't we ranking on Google Maps / locally?"
Before pitching SEO as a service, or kicking off a new SEO retainer (pairs with /seo-site-structure next).

NOT for: paid ads diagnosis (that's media-buyer / /meta-ads-analyze), or content writing (that's /seo-content).

What it produces

A Google Doc in Klijenti/<Client>/SEO/ titled "Local SEO Audit — <Client>", with:

Snapshot — business, location, site, primary services, current GBP signals seen.
Competition assessment — is this niche+location winnable? (uses the playbook's low-competition signals).
Technical audit — findings against the playbook checklist (indexability, speed signals, mobile, title/meta, NAP, internal linking, schema presence).
GBP gap analysis — what's missing/weak vs the GBP playbook (categories, description, services, posts, ACME Agencyws).
Core 30 / content gap — how far the site is from the Core 30 structure (topical + geographic relevance).
AI-visibility gap — is the business present in ChatGPT / AI Overviews / Bing (quick GEO read).
Prioritized 30/60/90-day plan — ordered actions, each tagged with the skill that executes it.

Step-by-Step

1. Parse the request

First positional = client/business name. Flags: --url, --location (city/region), --gbp (profile name), --dry-run (no Drive/Slack — print only).
If no URL and the client isn't in clients.json with a website, ask the user for the site URL and the target city. Don't guess.

2. Gather context (script)

node ACME Agency/scripts/seo_run.mjs --skill local-audit --mode gather \
  --client "<name>" --url "<url>" --location "<city>"

This prints a JSON CONTEXT block: resolved business facts (Slack channel, Drive folder), client notes, and the scraped site (homepage title/description/markdown + key pages). Read it.

3. Read the method

Read .claude/context/ai-seo/playbook/01-local-seo-audit.md in full (path is in the CONTEXT playbook field).
Skim .claude/context/ai-seo/playbook/00-field-notes.md for real-world gotchas (proximity ceiling, visible-address rule, "every URL needs a backlink", phone-answer-rate signal).
For the GBP / structure / AI-visibility gap sections, also glance at playbooks 02, 04, and 03 so the gaps are framed in those skills' terms.

4. Do the audit

Working from the scraped site + the playbook, write the deliverable as markdown (use #/##/###, bullet lists, and pipe tables). Be specific and evidence-based — cite what you actually saw on the site. Apply the playbook's rules literally:

Topical before geographic — flag if the site is chasing city terms without topical depth.
GBP proximity ceiling — if location targeting exceeds the GBP's real address reach, say so and set expectations.
Competition — give a clear winnable / hard / very-hard verdict with reasons.
Every recommendation must map to one of: /gbp-optimize, /seo-site-structure, /seo-content, /seo-schema, /seo-backlinks, /ai-visibility. Tag each action with its skill.
End with the 30/60/90 plan as a table: | Phase | Action | Skill | Expected impact |.

Write it to a temp file, e.g. ACME Agency/clients/<Client>/SEO/audit.md (create the folder) or /tmp/seo-audit.md.

5. Save + report (script)

node ACME Agency/scripts/seo_run.mjs --skill local-audit --mode save \
  --client "<name>" --in "<path to audit.md>" \
  --title "Local SEO Audit — <Client>" \
  --slack "*Local SEO Audit — <Client>*
Competition: <one-word verdict>. Top 3 priorities: <a>, <b>, <c>."

The script creates the Google Doc in the client's Drive /SEO folder and posts the summary + Doc link to the client Slack channel (one message — no progress chatter).
With --dry-run, append --dry-run: it prints what it would do and skips Drive/Slack.
Output the Doc link to the user immediately.

Guardrails

Recommend-only. Never edit the live site, GBP, or DNS. This skill diagnoses and plans.
No invented data. If you couldn't see the GBP or rank data, say "not verified" rather than asserting. Pull before hypothesizing (Diagnosis Protocol).
Croatian/Bosnian clients: any client-facing prose in the Doc follows the no-em-dash / no-AI-tell rules in root CLAUDE.md.