seo-meta-optimizer

A Claude Code skill that optimizes title tags and meta descriptions for SEO at scale.

What it does

Give it a website URL or CSV export and it will:

• Crawl your site — automatically extract all current titles and meta descriptions via sitemap or link crawling
• Fix title tags — shorten to ≤60 characters using grammar-aware truncation
• Fix meta descriptions — expand short ones and trim long ones to the 120-160 character sweet spot
• Detect grammar issues — trailing commas, dangling prepositions, unclosed parentheses, incomplete clauses
• Eliminate duplicates — find and differentiate pages sharing identical titles or metas
• Proper-case tech terms — PostgreSQL, AWS, Kubernetes, etc.
• Generate from context — create titles/metas for pages that have none, using URL path analysis

Install

Clone the repo and copy the skill files into your Claude Code commands directory:

Global install (available in all projects):

git clone https://github.com/djforge/seo-meta-optimizer.git
mkdir -p ~/.claude/commands
cp seo-meta-optimizer/skills/optimize-meta-tags/SKILL.md ~/.claude/commands/optimize-meta-tags.md
cat seo-meta-optimizer/skills/optimize-meta-tags/reference.md >> ~/.claude/commands/optimize-meta-tags.md

Per-project install (available only in a specific project):

git clone https://github.com/djforge/seo-meta-optimizer.git
mkdir -p .claude/commands
cp seo-meta-optimizer/skills/optimize-meta-tags/SKILL.md .claude/commands/optimize-meta-tags.md
cat seo-meta-optimizer/skills/optimize-meta-tags/reference.md >> .claude/commands/optimize-meta-tags.md

After installing, restart Claude Code. The /optimize-meta-tags command will be available.

Usage

From a website URL (crawls the site automatically):

/optimize-meta-tags https://example.com

Or from a CSV export:

/optimize-meta-tags path/to/your-meta-tags.csv

The skill will ask for:

1. Your brand name (e.g., "Acme Corp") for the title suffix
2. A one-line brand description for generating meta descriptions
3. Your website URL (if not already provided — fetched for brand voice and positioning context)
4. Any audit CSVs (nice-to-have — Ahrefs, Screaming Frog, etc. for organic traffic data and pre-flagged issues)

Output

A CSV file with columns:

Column	Description
URL	Page URL
Page_Type	website, blog, docs, learn, etc.
Organic_Traffic	From audit data (0 if unavailable)
Current_Title	Original title tag
Optimized_Title	New title tag
Title_Changed	YES or NO
Current_Meta	Original meta description
Optimized_Meta	New meta description
Meta_Changed	YES or NO

Plus optional split by section (website, blog, docs, learn) in a by_section/ subfolder.

Validation

Every run ends with a validation report targeting 0 issues across:

• Title length violations (>60 chars)
• Meta length violations (<120 or >160 chars)
• Grammar issues (30+ checks)
• Duplicate titles
• Duplicate meta descriptions

How it works

The optimizer uses a strategy chain for titles — trying progressively more aggressive approaches until one fits:

1. Full title + brand suffix
2. Apply shortenings (remove "Understanding", "Introducing", etc.) + brand
3. Grammar-aware truncation + brand
4. Drop brand suffix entirely
5. Unpack/remove parentheticals
6. Aggressive truncation at first colon or dash

Every truncation point is validated against grammar rules — it will never create a title ending with "and", "for", "the", a trailing comma, or an unclosed parenthesis.

License

MIT