# LLM SEO Audit

> Checks whether a page is quotable by generative assistants, not merely rankable. TRIGGER WHEN: the user asks to audit for AI search or answer engines, check visibility in Google AI Overviews / SGE, Perplexity, ChatGPT Search, Claude Search, or Bing Copilot, verify the AI-bot crawler allowlist, or check E-E-A-T, llms.txt, JSON-LD citation readiness, passage extractability, and prompt-injection hardening. DO NOT TRIGGER WHEN: the task is traditional organic / SERP ranking SEO (use /digital-marketing:seo-audit), or content tone and voice only (use the content-marketer agent).

- Skill: `acaprino/llm-seo-audit` (Agent Skill)
- Install (CLI): `npx skillmds@latest add acaprino/llm-seo-audit`
- Raw SKILL.md: https://api.skillmd.com/api/skills/acaprino/llm-seo-audit/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Marketing & Growth
- Author: acaprino (https://skillmd.com/u/acaprino)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/acaprino/llm-seo-audit

---


> Arguments: `<url or local path> [--focus <comma-separated: crawlers,eeat,schema,passages,injection | all>] [--strict-mode]`. Wherever `<arguments>` appears below, substitute the text the user typed after the skill name.

<!-- Generated by the Daodan compiler for codex. Edit the kernel, never this file. -->

# LLM SEO / Answer Engine Optimization Audit

Invokes the `llm-seo-optimize` agent to audit a site for answer-engine discoverability and citation-worthiness. Different from `/seo-audit` (traditional SERP ranking) -- this optimizes for getting **quoted inside an LLM-generated answer**.

## CRITICAL RULES

1. **Delegate to the `llm-seo-optimize` agent**. This command is a thin wrapper -- the agent owns the full 6-phase audit.
2. **Verify live, not just code**. If Playwright MCP is available, use it to confirm crawler access (`robots.txt` live fetch), JSON-LD presence, and rendered passage extractability.
3. **Write output to `.aeo-audit/` for persistence** so re-runs can diff against the baseline.
4. **Never fabricate crawler policies**. If `robots.txt` cannot be fetched, state the gap explicitly.
5. **Complementary to `/seo-audit`, not a replacement**. Traditional SEO and AEO require different optimizations -- flag the overlap, not duplication.

## Procedure

### 1. Parse arguments

- `<url or local path>`: required target -- live URL or path to static site / built HTML output
- `--focus`: restrict the audit to a subset of dimensions. Values are comma-combinable (`--focus schema,eeat`). Default: `all`, which runs every phase
  - `crawlers` -- the AI-bot robots.txt allowlist check
  - `eeat` -- E-E-A-T signal audit (authorship, citations, dates)
  - `schema` -- JSON-LD / Schema.org structured data, plus citation readiness
  - `passages` -- passage-level extractability (direct-answer paragraphs, tables, bullet lists)
  - `injection` -- prompt-injection hardening (hidden text, invisible CSS, comment payloads)
- `--strict-mode`: report-level severity escalation. Every Warning is raised to Critical and the report opens with an explicit `VERDICT: PASS` / `VERDICT: FAIL` line, FAIL when any Critical remains

### 2. Spawn the agent

Invoke `digital-marketing:llm-seo-optimize` with the target, focus, and strict-mode flag. The agent loads its own knowledge base and runs the 6-phase protocol.

### 3. Report

Agent writes `.aeo-audit/REPORT.md` with:

```
# AEO Audit -- <site/page> -- <date>

## Summary
- Pages audited: N
- Blocked from: <list of engines with robots.txt denies>
- E-E-A-T score: X/5
- Extractability score: X/5
- JSON-LD coverage: P%

## Critical findings
- [CRITICAL] <page:line / selector> <issue>

## Per-page findings
...

## Cross-cutting recommendations
...

## Tracking setup
- Analytics filters for AI-referral hostnames: chatgpt.com (legacy chat.openai.com), perplexity.ai, claude.ai, copilot.microsoft.com, gemini.google.com
- Weekly brand-mention query set for citation-share tracking
```

## Typical flow

```
# Full audit against a live URL
/llm-seo-audit https://example.com

# Focus on Schema + E-E-A-T only
/llm-seo-audit https://example.com --focus schema,eeat

# Static site audit
/llm-seo-audit ./dist/

# Strict grading: warnings become critical, report opens with PASS/FAIL
/llm-seo-audit https://example.com --strict-mode
```

## Complementary commands

Run alongside traditional SEO tooling for a complete picture:

| Command | Purpose | Overlap with AEO |
|---------|---------|------------------|
| `/seo-audit` | Technical SEO (Core Web Vitals, meta tags, sitemap, redirects, SERP ranking) | Low -- traditional ranking signals |
| `/content-strategy` | Conversion / CTA / tone / funnel audit | Medium -- answer-worthiness often correlates with clarity |
| `/ga4-audit` | GA4 + GTM + Consent Mode v2 verification | Measurement only -- AEO needs AI-referrer tracking set up in GA4 |
| **`/llm-seo-audit`** | **AEO / answer-engine optimization** | -- |

## Synergies

- Deep knowledge base -> `digital-marketing:llm-seo-optimize` agent (the actual worker)
- Structured data validation alongside broader checks -> `/digital-marketing:seo-audit`
- Measurement setup -> `digital-marketing:ga4-implementation-expert` agent (AI-referrer tracking)
- Humanizing AI-sounding copy to raise E-E-A-T -> `/text-humanizer:humanize-text` (text-humanizer plugin)
- Playwright-based live verification -> `playwright-skill`