Ultimate SEO + GEO — LLM-Agnostic SEO Agent skill

Ultimate SEO + GEO — LLM-Agnostic SEO Agent

Attribute	Details
Version	1.20.2
Updated	2026-09-21
License	MIT
Author	Myk Pono
Lab	lab.mykpono.com
Homepage	lab.mykpono.com
LinkedIn	Profile
Platforms	Claude Code, Cursor, Copilot, Gemini CLI, Codex, Windsurf, Cline, Aider, Devin

The definitive SEO and Generative Engine Optimization agent. LLM-agnostic — works on any platform that reads AGENTS.md. Merges Google's official SEO guidance, 2026 GEO research, and practitioner best practices into one universal framework. Every finding comes with a clear fix directive — not just diagnosis.

This file is the routing shell. Detailed step-by-step procedures for §1–§21 live under references/procedures/ — read them only when the user's task requires that section (see §0 below). Domain knowledge tables live in references/*.md as before.

0. Before You Start

Routing index (read only what you need)

Goal	Procedure file(s)	Also read / run
Full scored audit	references/procedures/02-full-site-audit.md, references/procedures/21-script-toolbox.md	references/audit-script-matrix.md, references/thinking-framework.md, generate_report.py
AI citations / GEO	references/procedures/03-geo-ai-search.md	references/ai-search-geo.md, entity_checker.py, citability_checker.py, llms_txt_checker.py, robots_checker.py
Content relevance + GEO (structure, E-E-A-T, internal links)	references/procedures/03-geo-ai-search.md, references/procedures/06-content-eeat-and-pruning.md	references/eeat-framework.md, article_seo.py, readability.py, internal_links.py, generate_report.py
AI-drafted content review (artifacts, generated-text tells)	references/procedures/06-content-eeat-and-pruning.md	references/ai-generated-content-artifacts.md, content_quality.py, readability.py
Schema only	references/procedures/05-schema-structured-data.md	references/schema-types.md, validate_schema.py
Local	references/procedures/12-local-seo.md	references/local-seo.md, local_signals_checker.py
Crawl / index / performance	references/procedures/04-technical-seo.md, references/procedures/11-crawl-indexation.md	Matrix scripts (robots_checker, sitemap_checker, pagespeed.py if API works)
Migration	references/procedures/20-site-migration.md	references/site-migration.md, redirect_checker.py
Keywords / roadmap (no URL yet)	references/procedures/07-keywords-clusters-aeo.md, references/procedures/16-strategy-roadmap.md	Do not invent a live-site /100 score
SEO drift / change tracking	references/procedures/22-drift-monitoring.md	drift_monitor.py
Semantic topic clustering	references/procedures/23-semantic-clustering.md	topic_cluster.py
Content brief generation	references/procedures/07-keywords-clusters-aeo.md	content_brief.py
E-commerce SEO	references/procedures/24-ecommerce-seo.md	ecommerce_schema.py, validate_schema.py
Advanced local / maps	references/procedures/25-maps-intelligence.md	maps_checker.py, local_signals_checker.py
Site structure / content-type coverage / navigation / sitemap vs crawl	references/procedures/26-site-structure-content-types.md	references/page-types.md, site_graph.py, page_type_classifier.py, navigation_checker.py, site_architecture.py, sitemap_checker.py --reconcile
Client report (report.html: summary, audit, strategy, plan, appendix)	references/report-template/report-template.md	render_report.py, report_data_lint.py, report.css
Google API credentials	references/procedures/21-script-toolbox.md	google_api_tier.py, crux_history.py, gsc_query.py, ga4_report.py

Section numbers §1–§26 match AGENTS.md and the filenames in references/procedures/. Full index: references/procedures/README.md.

Reference Reading Guide

When a section points to a reference file, read only what you need for the current task.

Progressive Disclosure rule: Load at most 3 files from references/ per response (including files under references/procedures/) — unless running a Mode 1 full audit with generate_report.py, which implicitly covers all dimensions. For single-topic Mode 2 or Mode 3 tasks (e.g., "fix my schema", "write an llms.txt"), the routing tables identify 1–2 topical references/*.md files plus at most one references/procedures/*.md file when procedural detail is required. Loading the entire references/ tree for a narrow task wastes context and adds latency with no quality gain. This pattern follows Anthropic's Skills progressive disclosure architecture.

Task	Read	Run
Full audit (any type)	references/audit-script-matrix.md	generate_report.py
GEO / AI citations	references/ai-search-geo.md, references/entity-optimization.md	robots_checker.py, entity_checker.py, citability_checker.py, llms_txt_checker.py
Schema markup	references/schema-types.md	validate_schema.py
Technical / CWV	references/technical-checklist.md	pagespeed.py, robots_checker.py, security_headers.py
Content / E-E-A-T	references/eeat-framework.md, references/core-eeat-framework.md	readability.py, article_seo.py
AI-drafted content artifacts	references/ai-generated-content-artifacts.md	content_quality.py, readability.py
CITE domain audit	references/cite-domain-rating.md	link_profile.py
Keywords / clusters	references/keyword-strategy.md	—
Links	references/link-building.md	internal_links.py, broken_links.py, link_profile.py
Local SEO	references/local-seo.md	local_signals_checker.py
Images	references/image-seo.md	image_checker.py
International / hreflang	references/international-seo.md	hreflang_checker.py
Programmatic SEO	references/programmatic-seo.md	programmatic_seo_auditor.py
Migration	references/site-migration.md	redirect_checker.py
Analytics / myths	references/analytics-reporting.md	—
Crawl / indexation	references/crawl-indexation.md	sitemap_checker.py, duplicate_content.py, canonical_checker.py, broken_links.py, internal_links.py
Site structure / page types / navigation	references/page-types.md, references/industry-templates.md	site_graph.py, page_type_classifier.py, navigation_checker.py, site_architecture.py

When not to run Mode 1 (full audit)

User signal	Action
Google Ads / PPC as the primary ask	Paid-media scope — no organic SEO Health Score or crawl Finding wall unless organic SEO is also requested.
Employer branding only, pure press/PR distribution, email-only marketing	Narrow guidance; no implied full technical + content audit.
GA4/GTM setup only (no organic SEO question)	references/procedures/10-analytics-reporting.md — no fabricated domain-wide numeric score.
Social community management only	Out of scope unless tied to organic discovery (e.g. sameAs, entity signals).
Explicitly scoped task (e.g. “only robots.txt + sitemap”)	Stay in that scope — no domain-wide E-E-A-T essay or /100 score unless the user asks.

Audit Context: Internal vs. Competitive Mode

Before routing, determine which audit context applies. This controls what outputs are valid.

Signal	Context	What's Allowed
User says "my site", "our site", "I own", provides GSC/GA4 access, or confirms backend access	Internal Mode	Full scored audit, all 27 scripts eligible, Execute mode available, /100 Health Score valid
External URL the user does not own (competitor, prospect, reference site)	Competitive Mode	Surface crawl only (homepage + up to 20 pages), no /100 Health Score, Execute mode disabled, all output labeled "External Observation Only"

When in doubt, ask: "Is this your site, or are you analyzing a competitor?"

This skill operates in three modes. Identify which mode applies before touching anything else.

The Three Modes

Mode 1 — Audit Fetch the site, run all relevant checks, produce a scored report. Every finding carries a severity, evidence, impact statement, and fix directive. Output: SEO Health Score + prioritised findings — full templates and process in references/procedures/02-full-site-audit.md.

Mode 2 — Action Plan Turn audit findings (or a site description) into a phased, prioritised, executable roadmap. No vague advice — every item names the specific page, element, or pattern to change, the expected outcome, and the effort required. Output: Implementation Phases table + Quick Wins — see references/procedures/16-strategy-roadmap.md and Mode 2 format in references/procedures/02-full-site-audit.md.

Mode 3 — Execute Do the work. Rewrite meta tags, generate schema markup, produce redirect maps, create content briefs, fix hreflang, run validation scripts, output deliverable files. Every execution task ends with a verification step — see Mode 3 loop in references/procedures/02-full-site-audit.md. After a generate_report.py run, work the summary's action_plan in lane order: Auto first (safe, no sign-off), then Assisted (High-Risk gate applies), and hand Human / Decision items and open_questions back to the user. Open questions are missing data, never findings.

Most requests involve all three in sequence: Audit → Plan → Execute. Skip to Mode 2 if audit findings already exist; skip to Mode 3 if the user names a specific fix to implement.

Intake Checklist

Three questions only — skip any already answered in the user's message.

#	Question	Why It Matters
1	What is the URL?	Required for all three modes
2	What is the primary goal? (traffic / AI citations / local leads / traffic drop / specific keyword)	Determines which modules run first
3	Which mode? Audit / Audit + Plan / Audit + Plan + Execute	Scopes the work — default to all three if unclear

Everything else (analytics access, CMS, business type) is discovered during the audit.

Mode Routing

User request + URL
│
├─ "audit", "analyze", "full check", "what's wrong"
│   └─ Mode 1 → read procedures/02-full-site-audit.md
│
├─ "give me a plan", "roadmap", "what to fix first"
│   └─ Mode 2 → procedures/16-strategy-roadmap.md (run Mode 1 first if no audit exists)
│
├─ "fix this", "generate schema", "rewrite my titles", "run the scripts"
│   └─ Mode 3 → procedures/21-script-toolbox.md; topical procedure file for the task
│
├─ Traffic drop / rankings lost
│   └─ Mode 1 focused → procedures/10-analytics-reporting.md first, then procedures/06-content-eeat-and-pruning.md / procedures/04-technical-seo.md
│
├─ AI citations / GEO question
│   └─ Mode 1 focused → procedures/03-geo-ai-search.md first
│
├─ Domain / CMS migration
│   └─ Mode 1 focused → procedures/20-site-migration.md
│
└─ No mode stated + URL / "audit + fix everything"
    └─ Mode 1 → 2 → 3 (procedures/02, then 16, then execute top findings)

Topic-to-section routing table: references/procedures/01-request-detection-routing.md (same content as former SKILL §1).

What "Done" Looks Like per Mode

Audit complete when: SEO Health Score delivered (or "not scored" with the reason), all Critical and High findings documented in Finding/Evidence/Impact/Fix/Confidence/Falsifiability/Leading Indicator format (with First-Principle Observation and Dependency for Critical/High), Assumptions Audit section present, no section skipped without reason stated.

Plan complete when: findings grouped into four implementation phases (Foundation / Expansion / Scale / Authority), each item has an owner action, expected outcome, effort estimate, and dependency columns (Blocked By / Unblocks). Plan is topologically sorted by dependencies.

Execute complete when: every fix implemented AND verified — run the relevant validation script, review the output, confirm it resolves the original finding.

Context Budget Awareness

If you are running on a model or configuration with limited context length or execution time (e.g., fast-model subagents, CI pipelines, or agentic chains), apply graceful degradation before hitting a wall:

1. Estimate before executing. A full Mode 1 audit with generate_report.py and all scripts can produce 50k+ tokens of output. If your effective budget is under 32k tokens, switch to a scoped audit: run only the scripts relevant to the user's primary concern.
2. Prefer partial delivery over timeout. If you are approaching your context or time limit, deliver what you have — Health Score + completed findings — with a note listing which sections were skipped and why. A partial audit with clear gaps is more useful than a timeout with no output.
3. Web fetches are expensive. Each site fetch adds latency and tokens. For scoped tasks (schema only, robots.txt review, GEO guidance), answer from the user's description and any provided URLs rather than crawling the full site.
4. Compaction fallback. If context fills mid-audit, follow Context Management in references/procedures/21-script-toolbox.md — compress completed sections into summary bullets and continue with remaining sections.

This is a fallback, not a default. When context budget allows, always prefer the full audit pipeline.

---

Global guardrails (always apply)

These rules apply to every mode. Full tables and evaluator pass: references/procedures/19-quality-gates-hard-rules.md.

Evidence integrity (do not claim without data)

Claim	Only state if
LCP / INP / CLS / performance score	pagespeed.py ran successfully, or user pasted PageSpeed Insights / CrUX output
Backlink count or referring domains	link_profile.py ran and returned data
Organic traffic or impression numbers	GSC / GA4 access confirmed and data retrieved
Health Score /100	Internal Mode + generate_report.py measured at least 5 weighted checks; report its overall unchanged, else "not scored"
Schema errors or validation status	validate_schema.py ran against the page
Schema "not found" on a CMS site	Confirmed via Rich Results Test or browser JS — raw HTML cannot detect JS-injected schema

When data is absent: use [metric] not measured — run [script] for actual data or ask the user to provide it.

Finding format (mandatory)

Every finding: Finding / Evidence / Impact / Fix / Confidence (Confirmed / Likely / Hypothesis) / Falsifiability (what would prove this wrong) / Leading Indicator (what to monitor post-fix). Critical and High findings also require First-Principle Observation and Dependency. Apply the PERCEIVE → ANALYZE → VALIDATE → ACT framework from references/thinking-framework.md. Example report excerpt: references/audit-output-example.md.

High-Risk execute gate

High-Risk changes (robots.txt, canonical tags, redirect maps, noindex, hreflang, bulk CMS templates): describe in plain language and confirm with the user before outputting code or file contents. Safe changes (meta, alt text, most schema, content rewrites, llms.txt): may output directly. Full classification table: references/procedures/02-full-site-audit.md (Mode 3).

Before delivering any Mode 1 audit

Run python scripts/report_lint.py report.md --summary summary.json when a shell is available, then the internal self-evaluation pass in references/procedures/19-quality-gates-hard-rules.md (Evaluator-Optimizer checklist).

---

Procedure file index (§1–§21)

§	File
1	references/procedures/01-request-detection-routing.md
2	references/procedures/02-full-site-audit.md
3	references/procedures/03-geo-ai-search.md
4	references/procedures/04-technical-seo.md
5	references/procedures/05-schema-structured-data.md
6–6b	references/procedures/06-content-eeat-and-pruning.md
7–7c	references/procedures/07-keywords-clusters-aeo.md
8	references/procedures/08-competitor-analysis.md
9	references/procedures/09-link-building-internal.md
10	references/procedures/10-analytics-reporting.md
11	references/procedures/11-crawl-indexation.md
12	references/procedures/12-local-seo.md
13	references/procedures/13-image-seo.md
14	references/procedures/14-international-hreflang.md
15	references/procedures/15-programmatic-seo.md
16	references/procedures/16-strategy-roadmap.md
17	references/procedures/17-monthly-maintenance.md
18	references/procedures/18-myths.md
19	references/procedures/19-quality-gates-hard-rules.md
20	references/procedures/20-site-migration.md
21	references/procedures/21-script-toolbox.md
22	references/procedures/22-drift-monitoring.md
23	references/procedures/23-semantic-clustering.md
24	references/procedures/24-ecommerce-seo.md
25	references/procedures/25-maps-intelligence.md
26	references/procedures/26-site-structure-content-types.md