neural-auditlisted

# Neural Portfolio Audit **First:** Read the file `~/.claude/skills/seo-references/core.md` before proceeding. Apply the core methodology and voice defined there to all output. **Also read:** `~/.claude/skills/seo-references/data-pull-patterns.md` for MCP tool parameter reference. ## Step 1: Pull Portfolio-Wide Data This skill has NO domain argument — it runs across the entire portfolio. Iterate through the MCP Routing Map in core.md and pull data for each domain that has live properties. **Semrush (PRIMARY — run for each domain):** 1. For each domain in the MCP Routing Map, call `mcp__semrush__execute_report` — Domain Overview (database=us). Captures: authority score, organic traffic estimate, keyword count. This is the most important data for the audit. **GA4 (ENRICHMENT — run for each property):** 2. For each GA4 property in the MCP Routing Map, call `mcp__google-analytics__run_report` with metrics=[sessions, bounceRate], dateRanges=[{startDate: "28daysAgo", endDate: "today"}]. Captures: real traffic + engagement. **GSC (ENRICHMENT — run for each property):** 3. For each GSC property, call `mcp__gscServer__get_search_analytics` with dimensions=[], rowLimit=1 (totals only). Captures: total clicks + impressions per property. **Note:** This skill makes many MCP calls. If rate limits or timeouts occur, collect as much data as possible and note which domains had incomplete data. ## Step 2: Analyze with Pareto Ruthlessness Using the pulled data + the Full 13-Domain Ros