safari-seo

name

safari-seo

description

Open-source agentic SEO skill for Tanzania safari, tour-operator, and travel sites. Works in Grok, Cursor, Claude Code, Codex, Copilot. Triggers on: safari SEO, tour operator SEO, Kilimanjaro/Serengeti/Ngorongoro ranking, TouristTrip schema, migration content, Northern Circuit keywords, local SEO Arusha Moshi, safari audit, hreflang, Core Web Vitals wildlife photos, or "perform SEO analysis on <url>". Ships 363 keywords, 15 sub-skills, 14 scripts. Outputs FULL-AUDIT-REPORT.md + ACTION-PLAN.md. Repo: github.com/cleven12/tanzania-tourism-seo

Safari SEO Skill (Agentic / Tanzania Tourism)

LLM-first SEO analysis skill purpose-built for Tanzania safari and tour-operator websites. Includes 15 specialized sub-skills, 7 specialist agents, and script-backed evidence collection for Kilimanjaro treks, Serengeti safaris, Ngorongoro crater trips, Northern Circuit packages, and Zanzibar add-ons.

<SKILL_DIR> = absolute path to this skill directory (the folder containing this SKILL.md).

Deterministic Trigger Mapping

Map common user wording to a fixed workflow:

User says	Route to
perform seo analysis on <url>	Single-URL full audit → resources/skills/safari-page.md
audit my safari site / full SEO check	Full site audit → resources/skills/safari-audit.md
optimize tour page / new safari package	resources/skills/safari-tour-page.md
serengeti page / kilimanjaro landing page	resources/skills/safari-destination.md
keyword research tanzania	resources/skills/safari-keywords.md
google business profile arusha	resources/skills/safari-local.md
schema for safari package	resources/skills/safari-schema.md
hreflang swahili english	resources/skills/safari-hreflang.md
competitor analysis safari	resources/skills/safari-competitors.md
AI search / ChatGPT citations	resources/skills/safari-geo.md
No explicit sub-skill + URL provided	Single-URL full audit with mandatory deliverables

For full/page audits, always produce:

• FULL-AUDIT-REPORT.md (detailed findings with evidence)
• ACTION-PLAN.md (prioritized fixes)

Available Commands

Command	Sub-Skill	Description
safari audit <url>	safari-audit	Full site audit with Tanzania scoring
safari page <url>	safari-page	Deep single-page analysis
safari content <url>	safari-content	E-E-A-T, depth, safari trust signals
safari technical <url>	safari-technical	Crawlability, CWV, indexing, security
safari schema <url>	safari-schema	TouristTrip, Product, TouristAttraction JSON-LD
safari local <url>	safari-local	GBP, citations, Arusha/Moshi local SEO
safari keywords <topic>	safari-keywords	Park-level keyword & seasonality research
safari plan <url>	safari-plan	Strategic content calendar & hub architecture
safari competitors <query>	safari-competitors	SERP gap analysis vs ranking operators
safari hreflang <url>	safari-hreflang	EN/DE/FR/SW international validation
safari images <url>	safari-images

---

Orchestration Logic

Step 1 — Identify the Task

Parse the user's request and read the matching resources/skills/safari-*.md file before executing.

• Full audit: safari-audit.md — crawl key pages (homepage, top tours, destination hubs, blog), delegate to agents
• Single page: safari-page.md — deep dive on one URL
• Tour launch: safari-tour-page.md + checklists/new-tour-page.md
• Destination hub: safari-destination.md + checklists/new-destination-page.md
• Strategic plan: safari-plan.md + seasonality from 04-tanzania-niche/seasonality.md

Step 2 — Load Tanzania Ground Truth

Always load references.md when citing fees, visas, park rules, or authority facts.

Fact type	Authoritative source	Never
National park fees	tanzaniaparks.go.tz (TANAPA)	Guess USD amounts
Ngorongoro fees	ncaa.go.tz	Conflate with TANAPA
Visa rules	immigration.go.tz	Outdated blog posts
Industry standards	tatotz.org (TATO)	Invent certifications
Porter ethics	kpap.org	Generic "responsible tourism"
Keyword volumes	04-tanzania-niche/keywords/*.csv bands	Fabricate exact MSV

Step 3 — Collect Evidence

Primary (LLM-first): Use read_url_content or direct page fetch for baseline reasoning.

Deterministic verification (when script execution is available):

# Fetch and parse page HTML
python3 <SKILL_DIR>/scripts/fetch_page.py <url> --output /tmp/safari-page.html
python3 <SKILL_DIR>/scripts/parse_html.py /tmp/safari-page.html --url <url> --json

# Tanzania-specific audit runner
python3 <SKILL_DIR>/scripts/seo-audit.py <url> --json

# Core Web Vitals (mobile-first — safari travelers book on phones)
python3 <SKILL_DIR>/scripts/pagespeed.py <url> --strategy mobile --json

# Robots, security, redirects
python3 <SKILL_DIR>/scripts/robots_checker.py <url>
python3 <SKILL_DIR>/scripts/security_headers.py <url>
python3 <SKILL_DIR>/scripts/redirect_checker.py <url>

# Schema validation against repo templates
python3 <SKILL_DIR>/scripts/validate_schema.py <file_or_url>
python3 <SKILL_DIR>/scripts/generate-schema.py --template touristtrip --input tour.yaml --json

# Sitemap analysis
python3 <SKILL_DIR>/scripts/sitemap_checker.py <url>
python3 <SKILL_DIR>/scripts/sitemap-gen.py --urls urls.txt --output sitemap.xml

# Images (hero photos dominate LCP on safari sites)
python3 <SKILL_DIR>/scripts/image_inventory.py <url> --json
python3 <SKILL_DIR>/scripts/image_weight_audit.py <url> --json

# Local SEO signals
python3 <SKILL_DIR>/scripts/local_seo_checker.py <url> --json

# Hreflang for multilingual operator sites
python3 <SKILL_DIR>/scripts/hreflang_checker.py <url> --json

# Readability and content depth
python3 <SKILL_DIR>/scripts/readability.py /tmp/safari-page.html --json
python3 <SKILL_DIR>/scripts/article_seo.py <url> --keyword "<target_keyword>" --json

# Broken links (critical on itinerary pages with many internal links)
python3 <SKILL_DIR>/scripts/broken_links.py <url> --workers 5

# AI search readiness
python3 <SKILL_DIR>/scripts/llms_txt_checker.py <url>
python3 <SKILL_DIR>/scripts/citation_readiness.py <url> --json

Do not use third-party mirrors (e.g., r.jina.ai) as primary evidence when direct fetch or bundled scripts are available.

If a script fails (network, timeout, missing dependency), report as environment limitation — not a confirmed site issue. Continue with LLM analysis at Likely confidence.

Step 4 — Delegate to Specialist Agents

For comprehensive audits, adopt the relevant agent role from resources/agents/:

Agent	File	Focus
Technical SEO	safari-technical.md	Crawlability, CWV, mobile, JS rendering
Content Quality	safari-content.md	E-E-A-T, experience signals, thin content
Schema Markup	safari-schema.md	TouristTrip, Product, TouristAttraction
Local SEO	safari-local.md	GBP, NAP, Arusha/Moshi citations
Performance	safari-performance.md	LCP heroes, INP on booking widgets
Verifier	safari-verifier.md	Deduplicate, validate evidence, suppress contradictions
Tanzania Niche	tanzania-niche.md	Migration timing, routes, fees, seasonality

Step 5 — Apply Repo Quality Gates

Resource	Path	Use for
Meta tag formulas	01-onpage-content/meta-tags-guide.md	Title/description patterns
Page architecture	01-onpage-content/page-architecture.md	Hub-and-spoke tour/destination structure
Internal linking	01-onpage-content/internal-linking.md	Blog → tour → destination flows
Content briefs	01-onpage-content/content-briefs/	Kilimanjaro, Serengeti, Ngorongoro templates
Schema templates	02-technical/schema/	JSON-LD with {{PLACEHOLDER}} syntax
Crawl checklist	02-technical/crawl-checklist.md	Indexing gates
CWV thresholds	02-technical/core-web-vitals.md	LCP/INP/CLS targets
Hreflang guide	02-technical/hreflang-i18n.md	EN/DE/FR/SW patterns
GBP guide	03-local/google-business-profile.md	Operator categories, service areas
Seasonality	04-tanzania-niche/seasonality.md	Content calendar timing
Search intent	04-tanzania-niche/search-intent-by-park.md	Info vs commercial routing
Glossary	04-tanzania-niche/glossary.md	Swahili terms, TANAPA/TATO/KPAP
Competitor patterns	04-tanzania-niche/competitor-patterns.md	SERP observation framework
Launch checklists	checklists/	Tour page, destination, monthly SEO

Step 6 — Verify Findings

python3 <SKILL_DIR>/scripts/finding_verifier.py --findings-json <raw_findings.json> --json

Use verified output for final report tables. Cross-check Tanzania-specific claims against references.md authorities.

Step 7 — Score and Report

Tanzania Safari Scoring Weights (Full Audit)

Canonical source of truth — defined here and mirrored in resources/skills/safari-audit.md.

Category	Weight	Rationale
Content Quality & E-E-A-T	22%	Trust-critical YMYL-adjacent travel queries
Technical SEO	18%	Mobile booking, crawlability, indexation
On-Page SEO	15%	Title/meta/H1 for competitive safari SERPs
Schema / Structured Data	15%	TouristTrip, Product, Offer rich results
Local SEO & Trust	12%	Arusha/Moshi operators depend on local + GBP
Performance (CWV)	8%	Hero images and booking widgets affect CWV
Image Optimization	5%	Wildlife photos are primary content
AI Search Readiness (GEO)	5%	Growing AI travel research channel

Score Interpretation

Score	Rating
90-100	Excellent
70-89	Good
50-69	Needs Improvement
30-49	Poor
0-29	Critical

Step 8 — Mandatory Deliverables

For safari audit, safari page, and generic perform seo analysis on <url> flows:

1. Create FULL-AUDIT-REPORT.md at audit start; update as evidence arrives
2. Create ACTION-PLAN.md at audit start; prioritize Critical → High → Medium → Low
3. List all generated artifact paths in the final response
4. Include Environment Limitations section if scripts were blocked
5. Flag any unverified fee/visa claims with ⚠️ VERIFY AT SOURCE

Report Severity Levels

• 🔴 Critical — Blocks indexing, violates structured data policy, or misstates fees/visas
• ⚠️ Warning — Optimization opportunity (fix within 1 month)
• ✅ Pass — Meets or exceeds standards
• ℹ️ Info — Not applicable or informational

---

Tanzania Page-Type Detection

Auto-detect page type to apply correct checks:

Signals	Page Type	Sub-Skill	Min Words
Price, duration, itinerary, "book now"	Tour package	safari-tour-page.md	800+
Park name, wildlife, best time, no price	Destination hub	safari-destination.md	1,200+
Route comparison, packing list, how-to	Informational blog	safari-content.md	1,500+
Company name, team, licenses, contact	Operator brand	safari-local.md	500+
/de/, /fr/, hreflang	International	safari-hreflang.md	—

Schema Templates (Repo)

Pre-built JSON-LD in 02-technical/schema/:

Template	Page Type
touristtrip.json + product-offer-tour.json	Multi-day safari/trek packages
touristattraction.json	Park/destination hubs
organization-travelagency.json	Operator footer sitewide
review-aggregaterating.json	First-party reviews only
breadcrumblist.json	All pages with URL depth
faqpage.json	FAQ sections (note: rich results restricted for commercial FAQ schema — use for content structure, not rich result expectation)

Generate programmatically:

python3 <SKILL_DIR>/scripts/generate-schema.py --template product-offer-tour --input tour.yaml

Validate before publish:

# Rich Results Test: https://search.google.com/test/rich-results
python3 <SKILL_DIR>/scripts/validate_schema.py templates/tour-page.html

---

Critical Rules (Tanzania Safari)

1. INP not FID — INP is the sole interactivity metric (FID removed September 9, 2024). Never reference FID.
2. Never fabricate fees — Cite tanzaniaparks.go.tz or ncaa.go.tz with retrieval date; use ⚠️ VERIFY if live fetch failed.
3. Never fabricate reviews in schema — AggregateRating must reflect real first-party reviews only.
4. FAQ schema is restricted — FAQPage rich results limited to government/health authority sites. Commercial safari sites may use FAQ content structure without expecting FAQ rich results.
5. HowTo schema is deprecated — Never recommend for Kilimanjaro packing lists.
6. JSON-LD only — Never recommend Microdata or RDFa.
7. Migration claims need dates — Serengeti migration timing varies annually; cite general patterns from 04-tanzania-niche/seasonality.md, avoid false precision ("crossing on July 15").
8. Kilimanjaro route facts — Seven official routes (Marangu, Machame, Lemosho, Shira, Rongai, Northern Circuit, Umbwe); success rates vary by operator — don't invent percentages without source.
9. White-hat only — Refuse PBNs, fake reviews, scraped aggregator content, AI doorway pages per park.
10. Location page limits — Warning at 30+ near-duplicate destination pages; hard stop at 50+.
11. LLM-first, script-backed — Always attempt scripts; never block report on single script failure.
12. Mandatory file artifacts — FULL-AUDIT-REPORT.md + ACTION-PLAN.md for audit flows.
13. Commercial vs informational routing — Route info queries to blogs (safari-content); commercial queries to tour pages (safari-tour-page) per 04-tanzania-niche/search-intent-by-park.md.

---

Dependencies

pip install -r <SKILL_DIR>/scripts/requirements.txt
# Optional: Playwright for visual analysis
pip install playwright && playwright install chromium

Local repo scripts (M4+): seo-audit.py, sitemap-gen.py, generate-schema.py — see scripts/README.md.