Check if audit-report.json exists in the project root.

If audit-report.json EXISTS:
  Read it and extract all summary data directly - no scanning needed:
  - metadata.framework → framework
  - summary.totalClickableElements, summary.withTracking, summary.withoutTracking
  - categorized.cta.total, categorized.form.total, categorized.nav.total,
    categorized.media.total, categorized.outbound.total
  - existingTracking.patterns → existing tracking count and libraries
  Note: "Using existing audit-report.json from gtm-analytics-audit"

If audit-report.json NOT FOUND:
  Note: "audit-report.json not found. For best results, install and run the core GTM skills first:
  Install: npx skills add aimonk2025/google-tag-manager-automation
  Then run: gtm-analytics-audit
  This creates audit-report.json which all paid skills use to skip codebase scanning."
  Proceeding with direct codebase scan as fallback.
  Check package.json for framework (React, Next.js, Vue, Shopify, WordPress, etc.)
  Use Glob to find component files.
  Use Grep to count:
  - Button/CTA elements (<button, id="cta_.*")
  - Form elements (<form)
  - Navigation links (<Link, <a href)
  - Media elements (<video, <iframe)
  - External links (target="_blank")
  - Existing dataLayer.push calls (already tracked)

Whether from audit-report.json or scan, infer business model:
- Cart/product/checkout patterns in CTAs/forms → Ecommerce
- Trial/pricing/feature patterns in CTAs → SaaS
- Form/contact/download patterns → Lead generation
- Article/blog/newsletter patterns → Media/content

[Source: audit-report.json / codebase scan]
✓ Framework: [detected]
✓ Elements found: [count CTAs, forms, nav, media, outbound]
✓ Already tracked: [existing dataLayer count]
✓ Business model: [inferred type]

Check audit-report.json existingTracking.patterns first for ad scripts.
If not found there, scan codebase:
- fbq, connect.facebook.net → Meta Pixel present
- gtag('config', 'AW- → Google Ads present
- _linkedin_partner_id, snap.licdn.com → LinkedIn present
- ttq, analytics.tiktok.com → TikTok present
- Existing GTM tags via gtm-implementation-log.json

Note detected platforms for use in Phase 4 recommendations.
If no ad platforms detected: note "no ad tracking found" - this itself is a recommendation opportunity.

=== Your Top 5 Tracking Opportunities ===

Site: [framework] | Business model: [type] | [X] trackable elements found

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

#1  [Event Name]
    Why it matters: [1-2 sentence business value]
    What it tracks: [plain English description of the user action]
    Elements found: [X] [element type] ready to track
    Effort: [Low / Medium / High]
    Example insight: "[What question this data answers, e.g., 'Which CTA gets the most clicks?']"

#2  [Event Name]
    [same structure]

#3  [Event Name]
    [same structure]

#4  [Event Name]
    [same structure]

#5  [Event Name]
    [same structure]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Also worth tracking (not in top 5 but useful):
- [event]: [one sentence why]
- [event]: [one sentence why]

Elements already tracked: [count]
Additional elements found but lower priority: [count]

--- What to Do Next ---

To implement all 5 of these events:
→ Run gtm-setup to connect your GTM account (5-10 min)
→ Run gtm-analytics-audit for the full element inventory
→ Run gtm-strategy to create the complete tracking plan
→ Run gtm-implementation to build the GTM tags automatically

Total estimated time with full automation: 1-2 hours

Want to see the full tracking plan (not just top 5)?
Run: gtm-strategy