Seo Toolkit

SEO Toolkit — 13 SEO Commands for AI Agents

Load this skill whenever you need to run any SEO task: technical audit, Core Web Vitals analysis, crawlability checks, schema.org validation, keywords, meta tags, content, images, reports, competitor comparison, or autonomous fixes.

The skill works in two modes:

• URL mode — fetch external site + robots.txt + sitemap.xml + key pages
• File mode — analyze a local project (pages/, app/, public/, framework configs)

---

Commands (13 slash-commands)

| Command | Description | Mode | Triggers | |---|---|---|---| | [seo-audit](commands/seo-audit.md) | Full technical audit: meta, headings, alt, broken links, sitemap, robots.txt, canonical, URL structure | URL / File | /seo-audit, seo audit, technical audit | | [seo-speed](commands/seo-speed.md) | Core Web Vitals (LCP/CLS/INP), blocking resources, unused CSS/JS, caching | URL / File | /seo-speed, core web vitals, site speed | | [seo-structure](commands/seo-structure.md) | Internal linking, click depth, orphan pages, URL hierarchy, breadcrumbs | URL / File | /seo-structure, site structure, URL structure | | [seo-crawl](commands/seo-crawl.md) | Crawlability: robots.txt, noindex, canonical, redirect chains, sitemap.xml | URL / File | /seo-crawl, crawlability, robots txt | | [seo-schema](commands/seo-schema.md) | JSON-LD schema.org: Product, Article, Organization, Breadcrumb, FAQ, validation | URL / File | /seo-schema, json-ld, schema org, structured data | | [seo-keywords](commands/seo-keywords.md) | Keyword density, cannibalization, LSI, long-tail, intent mapping | URL / File | /seo-keywords, keywords, keyword research | | [seo-meta](commands/seo-meta.md) | Title/description generation: lengths, uniqueness, CTR, Open Graph, Twitter Cards | URL / File | /seo-meta, meta tags, title description | | [seo-headings](commands/seo-headings.md) | H1-H6 hierarchy, order, keywords in headings, accessibility | URL / File | /seo-headings, headings, h1 h2 h3 | | [seo-content](commands/seo-content.md) | Thin content, duplicates, readability (Flesch-Kincaid), E-E-A-T signals | URL / File | /seo-content, content audit, content analysis | | [seo-images](commands/seo-images.md) | Alt text, WebP/AVIF, lazy loading, dimensions, srcset | URL / File | /seo-images, image seo, image optimization | | [seo-report](commands/seo-report.md) | Comprehensive scored report (7 dimensions, weights 100%), weekly action plan | URL / File | /seo-report, seo report, SEO summary | | [seo-compare](commands/seo-compare.md) | Competitor comparison: gaps, overlapping keywords, content gaps | URL | /seo-compare, competitor seo, competitor analysis | | [seo-fix](commands/seo-fix.md) | Autonomous fix agent: P1-P5 prioritization, diff-before-apply, safety rules | File | /seo-fix, seo fix, automated seo fixes |

---

Operating modes

URL Mode (default for external sites)

User: /seo-audit https://lovii.ru
Agent: fetch https://lovii.ru + robots.txt + sitemap.xml + key internal pages → run the audit

File Mode (for local projects)

User: /seo-audit file:///path/to/project
Agent: analyze files in pages/, app/, public/, configs like next.config.js, astro.config.mjs, .htaccess

---

Canonical patterns (for development/enrichment)

Full deep-dive — in references/canonical-patterns.md. Key canons:

• Dual URL/FS mode — every command works in both modes: by URL (https://...) or local files (file:///path), like Playwright test fixtures
• Weighted scoring — 7 dimensions with weights (seo-report): Meta 20%, Content 20%, Crawl 15%, Images 15%, Schema 10%, Perf 10%, Links 10% (Google Lighthouse model)
• P1–P5 prioritization — autonomous fixes prioritized by impact (like security advisories)
• Diff-before-apply — mandatory diff display before applying changes (like GitHub PR review)
• Safety rules — never change URLs, delete content, or touch logic without confirmation

---

Usage examples

# Full audit of lovii.ru
/seo-audit https://lovii.ru

# Core Web Vitals for a local Next.js project
/seo-speed file:///path/to/lovii_demo

# JSON-LD schema.org validation on a page
/seo-schema https://lovii.ru

# Keywords + cannibalization
/seo-keywords https://lovii.ru

# Comprehensive report with an action plan
/seo-report https://lovii.ru

# Compare with a competitor
/seo-compare https://lovii.ru https://competitor.com

# Autonomous fixes (file mode only)
/seo-fix file:///path/to/lovii_demo

---

Canonical analogues

Full deep-dive — in references/canonical-patterns.md. Key canons:

• Google Lighthouse / PageSpeed Insights — Core Web Vitals thresholds (LCP ≤ 2.5s, CLS ≤ 0.1, INP ≤ 200ms), scoring weights
• Google Search Central / Search Console — crawlability rules, sitemap.xml, robots.txt, canonical, rich snippets
• schema.org — JSON-LD vocabulary (Product, Article, Organization, Breadcrumb, FAQ) and required properties
• Playwright Test — dual-mode fixtures: URL + file system
• Screaming Frog / Ahrefs / Semrush — crawl patterns, keyword cannibalization detection, content gaps
• GitHub Security Advisories — P1–P5 prioritization and diff-before-apply safety

---

Files

• SKILL.md — this file
• skill.json — manifest
• references/canonical-patterns.md — canonical patterns deep dive
• commands/seo-audit.md — full technical audit
• commands/seo-speed.md — Core Web Vitals
• commands/seo-structure.md — site structure & internal linking
• commands/seo-crawl.md — crawlability & robots.txt
• commands/seo-schema.md — JSON-LD schema.org
• commands/seo-keywords.md — keywords, density, cannibalization
• commands/seo-meta.md — title/description/OG/Twitter
• commands/seo-headings.md — H1-H6 hierarchy
• commands/seo-content.md — thin content, duplicates, readability
• commands/seo-images.md — alt, WebP/AVIF, lazy loading
• commands/seo-report.md — weighted scoring report
• commands/seo-compare.md — competitor comparison
• commands/seo-fix.md — autonomous fix agent
• scripts/seo_toolkit.py — helper script (HTML parser, keyword density counter, JSON-LD validator)

---

Installation

# For opencode
cp -r skills/seo-toolkit ~/.config/opencode/skills/

# For other agents
# Copy the skill folder to your skills directory

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: bestdeejay-design
• Source: bestdeejay-design/agent-skills
• License: MIT
• Homepage: https://bestdeejay-design.github.io/agent-skills/

Install and usage instructions live in the source repository linked above.