skills.rest
KristjanHS
KristjanHSCOMMUNITY
Explore all skills by this author →
GitHub

llm-security-auditor

Audit LLM skills and plugins for prompt injection and data exfiltration threats.

Updated Jun 19, 2025
Search Tags:#security#vulnerability#audit#llm#plugin#prompt injection
One-click install
npx skills add https://github.com/KristjanHS/hands-on-llm --skill llm-security-auditor
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llm-security-auditor
Source: https://github.com/KristjanHS/hands-on-llm/tree/main/skills/llm-security-auditor
Command: npx skills add https://github.com/KristjanHS/hands-on-llm --skill llm-security-auditor

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill identifies and reports on security vulnerabilities within LLM skills, plugins, or agent definitions, protecting against prompt injection and other malicious attacks.

Core Features & Use Cases

  • Comprehensive Auditing: Scans provided content (GitHub URLs, uploads, local paths, zip archives) for known threat categories.
  • Detailed Reporting: Generates a structured report outlining findings, severity, and explanations.
  • Sanitization Offer: Offers to automatically remove or neutralize identified threats from local files with user confirmation.
  • Use Case: Before deploying a new AI plugin, use this Skill to perform a security audit to ensure it doesn't contain hidden malicious instructions or vulnerabilities.

Quick Start

Use the llm-security-auditor skill to scan the GitHub repository at https://github.com/example/my-skill.

Frequently Asked Questions about llm-security-auditor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I audit an LLM plugin for prompt injection vulnerabilities?▼

To audit an LLM plugin for prompt injection vulnerabilities, provide the skill's GitHub URL, local path, or zip archive to scan for known threat categories and receive a detailed security report.

What is the best way to scan agent definitions for malicious code execution threats?▼

Scanning agent definitions for malicious code execution threats involves analyzing the provided files or repository links to identify security risks and generate a structured threat severity report.

Can I use this security audit skill to check a GitHub repository URL?▼

Yes, you can use this security audit skill to check a GitHub repository URL. The tool directly analyzes content fetched from provided GitHub links to detect data exfiltration and other malicious instructions.

How do I sanitize local files after identifying security threats in an AI skill?▼

To sanitize local files after identifying security threats in an AI skill, you can accept the optional sanitization offer generated at the end of the audit to automatically neutralize or remove the identified vulnerabilities.

What types of data sources can I scan when auditing LLM skills for security?▼

When auditing LLM skills for security, you can scan data sources including GitHub URLs, uploaded files, local file paths, and zip archives to detect malicious code execution and prompt injection.

Related Skills

Explore More Software Engineering Skills→
nicepkgnicepkg
2

skill-downloader

Download and install Claude Code skills from GitHub, archives, and URLs.

Official
Intermediate
zhangzhang-111-izhangzhang-111-i
2

skill-security-auditor

Scan AI agent skills for security vulnerabilities before installation.

Community
Advanced
ngquaanghuyngquaanghuy

skill-security-auditor

Analyze AI agent skill files for command injection, data exfiltration, and prompt injection.

Community
Advanced
hoanghn61hoanghn61

ai-threat-testing

Test LLM applications for OWASP LLM Top 10 vulnerabilities.

Community
Advanced