skill-trust-auditor

Community

Audit ClawHub skills for security risk.

AuthorWanli-Lee
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This Skill helps you prevent supply-chain and prompt-injection style security incidents by checking a ClawHub skill’s content for dangerous patterns before you install and run it.

Core Features & Use Cases

  • Automated security scanning: Fetches the target skill’s SKILL.md and referenced scripts, then runs 52 regex-based checks against known attack vectors.
  • Trust Score with explainable findings: Produces a 0–100 Trust Score plus a list of flagged risk items with exact file locations and matched snippets.
  • Optional LLM-as-judge mode: Uses Anthropic Claude Haiku as an advisory reviewer for ambiguous curl intent when an API key is available.
  • Safety-focused decision support: Converts scan results into clear verdicts (SAFE / INSTALL WITH CAUTION / RISKY / DO NOT INSTALL) and actionable next steps.

Quick Start

Ask OpenClaw to install the skill-trust-auditor skill, then tell your agent: Audit steipete/some-skill before I install it.

Dependency Matrix

Required Modules

python3requestsanthropic

Components

scriptsreferences

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: skill-trust-auditor
Download link: https://github.com/Wanli-Lee/CUA-Claw-Harness/archive/main.zip#skill-trust-auditor

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.