42-meta-optimizer

name: 42-meta-optimizer description: > AI-powered bulk meta description analysis: score quality on 4 criteria (0-40), rewrite to length constraints (desktop ≤160 chars, mobile ≤130 chars), and generate featured snippet "At a glance" summaries. Accepts single URLs, CSVs with URLs, or Screaming Frog Internal:HTML exports. No Python scripts — this is fully AI-based analysis and rewriting. Use when user says "meta description", "meta optimizer", "meta rewrite", "snippet", "featured snippet", "meta grade", "meta score", "meta quality", "SERP snippet", "description optimization", "meta bulk", "rewrite descriptions". version: 1.0.0 tags: [seo, meta-description, copywriting, featured-snippet, serp, ai-rewrite, bulk] allowed-tools:

• Read
• Write
• Bash
• WebFetch
• Grep
• Glob metadata: filePattern:
  • "**/meta*.csv"
  • "**/internal_html*.csv"
  • "**/descriptions*.csv"
  • "**/META-OPTIMIZER.md" bashPattern:
  • "meta.optim"
  • "meta.desc"
  • "meta.rewrite"

Meta Optimizer — AI-Powered Meta Description Analysis & Rewriting

Purpose

Analyze, score, rewrite, and optimize meta descriptions at scale using AI. No external scripts required — Claude performs all grading, rewriting, and snippet generation directly. Works with single URLs, CSV files, or Screaming Frog Internal:HTML exports.

Commands

/42:meta-optimizer grade <url-or-csv>          # Score meta descriptions (0-40)
/42:meta-optimizer rewrite <url-or-csv>        # AI rewrite to length constraints
/42:meta-optimizer snippet <url-or-csv>        # Generate featured snippet summaries
/42:meta-optimizer full <url-or-csv>           # All three: grade + rewrite + snippet

Input Formats

Single URL

Provide a URL directly. Claude fetches the page with WebFetch and extracts the existing meta description, title, H1, and page content for context.

CSV File

A CSV with at minimum a URL column. Optional columns:

• Meta Description or Description 1 — existing description text
• Title 1 or Title — page title
• H1-1 or H1 — primary heading
• Word Count — page word count

Screaming Frog Internal:HTML Export

Standard SF export with columns: Address, Title 1, Meta Description 1, Meta Description 1 Length, Meta Description 1 Pixel Width, H1-1, Word Count, Status Code. Read the CSV with the Read tool and parse accordingly.

Workflow

Step 1: Parse Input

1. Single URL: Use WebFetch to retrieve the page. Extract <title>, <meta name="description">, <h1>, and the first 500 words of body text.
2. CSV / SF Export: Use Read to load the file. Identify column mapping:
   • URL column: Address or URL or first column containing URLs
   • Description column: Meta Description 1 or Meta Description or Description
   • Title column: Title 1 or Title
   • H1 column: H1-1 or H1
3. For CSVs with many rows, process in batches of 25.

Step 2: Grade Meta Descriptions

Score each meta description on 4 criteria, each 0-10 points, for a total of 0-40.

Rubric

Score Interpretation

Step 3: Rewrite Meta Descriptions

For each description scoring below 30, generate optimized rewrites following these constraints:

Desktop Version (primary)

• Max 160 characters (hard limit)
• Target pixel width: ≤990 pixels (approximate: avg 6.2px per character)
• Structure: [Benefit/Hook] + [Core value] + [CTA or closing benefit]
• Must include the primary keyword naturally (infer from title + H1)
• End with a period or CTA verb phrase
• No quotes, no ALL CAPS, no excessive punctuation

Mobile Version

• Max 130 characters (hard limit)
• Tighter version of desktop — same message, fewer words
• Prioritize the single strongest benefit
• Still includes primary keyword

Rewrite Rules

1. Preserve brand voice: Match the tone of the existing description and page content (formal, casual, technical, friendly)
2. Primary keyword: Identify from title + H1 + URL slug. Place in first half of description when possible.
3. No keyword stuffing: Keyword appears once, maximum twice if natural.
4. End strong: Close with a CTA verb (discover, learn, explore, get, try, see, find, start) or a concrete benefit statement.
5. Front-load value: Most important information in first 80 characters (visible in all SERP truncation scenarios).
6. Avoid cliches: No "In this article", "Click here to", "Welcome to", "Looking for", "Read on to find out".

Step 4: Generate Featured Snippet Summaries

Create an "At a glance" summary optimized for Google Featured Snippet extraction.

Snippet Format

**At a glance:**
- [Bullet 1: Key fact or definition]
- [Bullet 2: Core benefit or primary answer]
- [Bullet 3: Supporting detail or statistic]
- [Bullet 4: Secondary benefit or scope] (optional)
- [Bullet 5: Action or next step] (optional)

Snippet Rules

• 3-5 bullets per page
• 40-60 words total (strict range)
• Factual and declarative: No hedging ("might", "could", "it depends"), no questions, no subjective claims
• Self-contained: Each bullet makes sense on its own
• Front-load answers: If the page answers a question, the first bullet should contain the direct answer
• Use specific numbers: Prefer "saves 3 hours per week" over "saves time"
• No brand fluff: Focus on information, not marketing
• Optimized for extraction: Google's featured snippet algorithm favors concise, structured, factual content in list or paragraph format

Step 5: Generate Report

Compile all results into META-OPTIMIZER.md.

Output Format

Write to META-OPTIMIZER.md in the current working directory (or specified output path).

# Meta Description Optimization Report

**Generated**: [date]
**Source**: [URL / filename]
**Pages analyzed**: [N]
**Average score**: [X/40]

---

## Grade Summary

| # | URL | Current Description | Score | Emotional | Benefit | Active | CTA | Rating |
|---|-----|-------------------|-------|-----------|---------|--------|-----|--------|
| 1 | /page-1 | "Current meta..." | 24/40 | 6 | 7 | 5 | 6 | Good |
| 2 | /page-2 | "Current meta..." | 12/40 | 3 | 4 | 2 | 3 | Below Avg |
| ... | | | | | | | | |

### Score Distribution
- Excellent (31-40): [N] pages ([%])
- Good (21-30): [N] pages ([%])
- Below Average (11-20): [N] pages ([%])
- Poor (0-10): [N] pages ([%])

---

## Rewrites

### /page-1
**Current** (score: 24/40):
> "Current meta description text here."

**Desktop rewrite** (158 chars):
> "Rewritten meta description with benefit-first structure and strong CTA."

**Mobile rewrite** (127 chars):
> "Shorter version with core benefit and CTA."

**Changes made**: Added power word "proven", led with user benefit, replaced passive "is designed" with active "delivers", added CTA "Get started today."

---

### /page-2
**Current** (score: 12/40):
> "Current weak meta description."

**Desktop rewrite** (155 chars):
> "Optimized description here."

**Mobile rewrite** (128 chars):
> "Mobile version here."

**Changes made**: [specific changes explained]

---

## Featured Snippets

### /page-1
**At a glance:**
- [Bullet 1]
- [Bullet 2]
- [Bullet 3]

*Word count: [N] words*

---

### /page-2
**At a glance:**
- [Bullet 1]
- [Bullet 2]
- [Bullet 3]
- [Bullet 4]

*Word count: [N] words*

---

## Implementation Checklist

- [ ] Update meta descriptions in CMS / codebase
- [ ] Verify character counts after CMS rendering (some CMS add suffix)
- [ ] Test SERP preview with Google SERP Simulator
- [ ] Add featured snippet content to page body (above the fold or in FAQ section)
- [ ] Re-crawl with Screaming Frog to verify changes
- [ ] Monitor CTR changes in Google Search Console (2-4 week window)

Cross-References

Tips

• Batch processing: For sites with 100+ pages, prioritize pages with highest impressions (from GSC) or those ranking in striking distance (positions 4-20).
• A/B testing: Use Google Search Console performance data to measure CTR impact of rewritten descriptions after 2-4 weeks.
• Character vs pixel width: Google truncates by pixel width (~990px for desktop), not character count. 160 chars is a safe proxy, but characters like "W" and "M" consume more pixels than "i" and "l".
• AI provider: This skill uses Claude's own capabilities for grading and rewriting. For alternative providers, pipe the rubric and constraints to OpenAI or another API via Bash.