amazon-catalog-auditor

Amazon Catalog Auditor - OpenClaw Skill

[!WARNING] Deprecated wrapper repo as of April 20, 2026. Active development now lives in amazon-catalog-cli. Please open issues and pull requests there instead of this repository.

OpenClaw skill for auditing Amazon Category Listing Reports

This skill wraps the amazon-catalog-cli tool to provide automated CLR analysis within OpenClaw workflows.

What It Does

Enables OpenClaw agents to audit Amazon CLRs with natural language commands:

• "Audit this CLR"
• "Check for missing attributes"
• "What RUFUS issues do I have?"
• "Analyze my catalog health"

The skill automatically:

• Runs catalog queries
• Parses results
• Presents prioritized insights
• Provides actionable recommendations

New in CLI v2.0.0 — Agent-First Redesign:

Note: v2.0 of amazon-catalog-cli now ships with a built-in SKILL.md and an MCP server (catalog mcp). For MCP-compatible clients (Claude Desktop, CLR Pro, etc.), the CLI itself is now a first-class agent surface — no separate skill wrapper needed. This OpenClaw skill remains useful for OpenClaw-specific workflows.

• MCP server - catalog mcp exposes catalog_scan, catalog_check, catalog_list_queries, catalog_schema as MCP tools
• JSON/stdin input - catalog scan --json '{...}' or piped via --stdin
• Schema introspection - catalog schema --format json for auto-discovery
• Field masks - --fields sku,severity,details to reduce output size
• Pagination - --limit and --offset for large catalogs
• NDJSON streaming - --format ndjson for line-by-line output
• Input validation - Rejects path traversal, injection, malformed input
• Backward compatible - All v1.x commands work unchanged

v1.3.0:

• Marketplace detection - Auto-detects US/CA/UK/DE/etc from CLR metadata
• Bullet awareness checks - Flags soft violations (excessive caps, problematic chars)
• 12 total queries

v1.2.0:

• Bullet point compliance - Comprehensive validation against Amazon's official requirements
• Prohibited content detection - Finds banned chars, emojis, placeholder text, prohibited claims
• Formatting validation - Checks capitalization, length, punctuation rules

v1.1.0:

• RUFUS tier scoring - Each SKU gets a health rating (Good/Fair/Weak/Critical)
• Catalog-wide summary - Overall score + distribution stats
• FBM duplicate filtering - Auto-skips redundant FBM/MFN SKUs (cleaner reports)

Installation

This wrapper is no longer the primary distribution target. Install and use amazon-catalog-cli directly for the maintained implementation.

Prerequisites

1. OpenClaw installed and running
2. amazon-catalog-cli installed:

   pip install amazon-catalog-cli
   

Install Skill

1. Download amazon-catalog-auditor.skill
2. Add to your OpenClaw skills directory
3. Restart OpenClaw

Or via OpenClaw CLI:

# Coming soon: OpenClaw skill install command

Usage

Basic Audit

You: Audit this CLR
[attach CLR file]

Agent: 🔍 CLR Audit Results

📊 Summary:
- 47 issues found
- 23 SKUs affected

🔥 Top Priorities:
1. Missing Brand (12 SKUs)
2. RUFUS Bullets scoring low (8 SKUs)
3. Titles exceeding 200 chars (3 SKUs)

💡 Recommendations:
- Add brand to 12 SKUs
- Rewrite Bullet 1 for 8 products
- Trim 3 titles under 200 characters

Specific Checks

You: Check my RUFUS bullet optimization

Agent: 📝 RUFUS Bullet Analysis

15 SKUs scoring below 4/5

Common Issues:
1. Bullet 1 missing Hero Benefit (8 SKUs)
2. Bullet 2 missing target audience (5 SKUs)
3. No specific data points (12 SKUs)

Top 3 to fix:
- ABC-123: Score 2/5 (no benefits)
- XYZ-456: Score 2.5/5 (vague language)
- DEF-789: Score 3/5 (needs specifics)

Export for Clients

You: Generate a report I can send to my client

Agent: Created client-audit.csv with all 47 issues.
Summary report ready to share.

Features

12 Built-in Checks

1. Missing Attributes - Required fields
2. Missing Any Attributes - Required + conditional
3. Long Titles - Over 200 characters
4. Title Prohibited Chars - Invalid characters
5. RUFUS Bullets - AI optimization scoring
6. Prohibited Chars - Invalid characters in title/brand
7. Bullet Prohibited Content - Detects banned chars, emojis, claims, placeholders (v1.2.0)
8. Bullet Formatting - Validates capitalization, length, punctuation (v1.2.0)
9. Bullet Awareness - Soft violations: excessive caps, problematic special chars (NEW v1.3.0)
10. Product Type Mismatch - Type/keyword alignment
11. Missing Variations - Variation opportunities
12. New Attributes - Unused template fields

Intelligent Analysis

• Prioritization - Critical issues first
• Context-aware - Understands Amazon requirements
• Actionable - Specific fix recommendations
• Export options - JSON, CSV, or human-readable

Examples

Example 1: Full Audit

User: "Audit this CLR and prioritize what needs fixing"

Agent analyzes and responds:

🔍 PlantFusion CLR Audit

📊 Summary: 47 issues, 23 SKUs

🔥 Critical (Fix Now):
• Missing Brand (12 SKUs) - Required field
• Prohibited characters (3 SKUs) - Blocks listing

⚡ Important (Fix Soon):
• RUFUS scores <3 (8 SKUs) - Poor conversion
• Missing conditional fields (15 SKUs) - Product type specific

💡 Optimize (When Time Allows):
• RUFUS scores 3-4 (12 SKUs) - Good, can be great
• Variation opportunities (4 SKUs) - Bundle related products

Quick wins: Fix brand + prohibited chars = 15 issues resolved

Example 2: RUFUS Deep Dive

User: "Why are my bullet points scoring low?"

Agent explains:

📝 RUFUS Analysis for 8 low-scoring SKUs

Position-specific issues:

Bullet 1 (Hero Benefit):
❌ "Premium quality hair gel with amazing hold"
✅ "Reduce frizz by 80% and add lasting shine"

Bullet 2 (Target Audience):
❌ "Great for styling"
✅ "Ideal for curly or wavy hair, perfect for all-day hold"

Common problems:
• Vague marketing (5 SKUs) - "premium", "amazing", "best"
• Missing data points (7 SKUs) - Add %, oz, count, time
• Too short (3 SKUs) - Expand with specifics

Want me to rewrite examples for your top 3 SKUs?

Requirements

• OpenClaw 2024.2+
• amazon-catalog-cli 2.0.0+ (recommended — includes MCP server, schema introspection, field masks)
• Python 3.10+

How It Works

1. User uploads CLR or provides file path
2. Skill runs catalog scan command
3. Parses JSON output (now includes marketplace metadata)
4. Analyzes by severity and priority
5. Presents human-readable insights
6. Offers export options if needed

The skill uses the CLI tool under the hood, so all 12 queries stay up-to-date automatically when the CLI is updated.

Security & Safety

This skill is designed with security in mind:

What It Does ✅

• Read-only operations - Only reads CLR files from your workspace
• Deterministic parsing - Pure Python logic, no AI model calls
• Local execution - All processing happens on your machine
• Output files only - Creates reports in workspace, nothing else
• Open source - Full code transparency

What It Doesn't Do ❌

• No network calls - Doesn't phone home or transmit data
• No credential storage - Doesn't ask for or store API keys
• No system modifications - Doesn't touch files outside workspace
• No external dependencies - Only uses amazon-catalog-cli (also open source)
• No telemetry - Zero tracking or analytics

Skill Safety Best Practices

When evaluating any OpenClaw skill (including this one):

1. Review the code - Check SKILL.md and scripts/ folder
2. Verify dependencies - Fewer is better, inspect what gets installed
3. Check the source - Install from trusted developers/official repos
4. Understand permissions - Know what file access the skill needs
5. Monitor behavior - Watch wh