EEAT OpenClaw Skill Audit

AI Agent Skill Quality Assurance — This skill adapts the CORE-EEAT framework to evaluate OpenClaw Skills, ensuring they deliver meaningful utility while maintaining security and reliability.

Skill Overview

OpenClaw Skills are modular capability extensions for AI agents, defined by SKILL.md files with YAML frontmatter and prompt instructions. This skill evaluates skill quality through 80 standardized criteria across 8 core dimensions, generating comprehensive audit reports including utility scores, security assessments, and actionable improvement recommendations.

Core Transformation:

• From: Install Skills blindly → Hope they work
• To: Systematic vetting → Data-driven skill selection

OpenClaw Skill Structure

Every OpenClaw Skill consists of:

my-skill/
├── SKILL.md          # Core definition (YAML + Markdown instructions)
├── scripts/          # Optional executable scripts
│   └── main.py
└── references/       # Optional configuration and resources
    └── config.json

Key Components:

• YAML Frontmatter: Skill metadata (name, description, version, dependencies, gates)
• Prompt Instructions: How the AI should use this skill
• Scripts: Optional executable code for complex operations
• Gating Mechanism: Conditional activation (bins, env, os checks)

Applicable Scenarios

Use this skill when users request:

Skill Selection

• "Evaluate this skill before installing"
• "Compare two skills for the same task"
• "Which skill should I use for X?"

Security Vetting

• "Is this skill safe to run?"
• "Scan for security vulnerabilities"
• "Check permission boundaries"

Skill Development

• "Audit my skill for quality issues"
• "How to improve my skill's documentation?"
• "What security best practices am I missing?"

Skill Maintenance

• "Review installed skills for quality"
• "Identify deprecated or risky skills"
• "Prioritize skill updates"

Core Capabilities

This skill can:

1. Complete 80-Item Audit: Score each CORE-EEAT item adapted for OpenClaw Skills

2. Utility Scoring: Evaluate task completion efficiency and comparative value

3. Security Assessment: Three-level security evaluation (Pass/Caution/Risk)

4. Gating Validation: Check conditional activation requirements (bins, env, os)

5. Veto Item Detection: Flag critical security violations (command injection, data leakage)

6. Priority Ranking: Identify top 5 improvements by impact

7. Comparative Analysis: Compare skills for same use case

Skill Categories

This skill supports 6 OpenClaw Skill types, each with different evaluation priorities:

Productivity Skills

• Definition: Gmail, Calendar, Google Drive, Microsoft Office integration
• Focus: Task completion accuracy, API reliability, error handling
• Weights: C: 30% | R: 25% | Exp: 20% | Ept: 15% | O: 5% | E: 0% | A: 5% | T: 0%