skill-security-audit

Comprehensive security risk assessment tool for ClawHub skills. Detect prompt injection, malicious scripts, supply chain attacks, credential theft, and other threats before installation — informed by real-world attack cases like the ClawHavoc campaign.

Features

• 60+ malicious pattern detection — covers RCE, Base64 obfuscation, reverse shells, credential theft, persistence mechanisms, prompt injection, and more
• Two scan modes — scan a local skill directory or download directly from ClawHub via --slug
• Zero external dependencies — pure Python standard library (urllib, zipfile, re, hashlib)
• Structured JSON report — severity-classified findings, file integrity inventory with SHA-256 hashes
• Threat knowledge base — reference documentation based on the ClawHavoc supply chain attack

Quick Start

Scan a local skill directory

python3 scripts/scan_skill.py /path/to/skill-directory

Scan a skill from ClawHub (not yet installed)

python3 scripts/scan_skill.py --slug stock-price-query

Scan a specific version

python3 scripts/scan_skill.py --slug stock-price-query --version 1.0.3

Usage as OpenClaw Skill

This project is designed to be used as a skill in OpenClaw. Once installed, the AI agent will:

1. Determine the skill source (local path, slug, or ClawHub URL)
2. Run the automated static analysis scanner
3. Perform expert review of SKILL.md for prompt injection and hidden directives
4. Review all scripts and code files for obfuscation, network behavior, and sensitive access
5. Generate a structured security assessment report with risk rating

Report Format

The scanner outputs a JSON report containing:

• scan_metadata — timestamp, tool version, scanned path
• summary — total findings count, breakdown by severity (CRITICAL / HIGH / MEDIUM / LOW)
• findings — each finding with severity, category, message, file, line number, and matched text
• file_inventory — SHA-256 hashes and sizes of all files for integrity verification

Risk Rating Criteria

Level	Description
CRITICAL	Active exploit patterns (RCE chains, credential theft, reverse shells, confirmed prompt injection)
HIGH	Dangerous patterns that could be weaponized (curl|bash, eval/exec, sensitive path access)
MEDIUM	Suspicious patterns with both legitimate and malicious use cases
LOW	Minor informational findings (non-standard URLs, code quality markers)
SAFE	No findings or only LOW-severity items with clear legitimate purpose

Project Structure

skill-security-audit/
├── SKILL.md                            # OpenClaw skill definition & workflow
├── README.md                           # This file
├── CHANGELOG.md                        # Version history
├── scripts/
│   └── scan_skill.py                   # Automated static analysis scanner
└── references/
    └── threat_knowledge_base.md        # Threat intelligence knowledge base

Requirements

• Python 3.10+
• No external packages required

License

MIT