skills.rest
cmndcntrlcyber
cmndcntrlcyberCOMMUNITY
Explore all skills by this author →
GitHub

hunt-llm-ai

Detect and mitigate AI and LLM vulnerabilities like prompt injection and exfiltration.

3|Updated Nov 12, 2025
Search Tags:#AI security#prompt injection#LLM security#exfiltration#agent misuse
One-click install
npx skills add https://github.com/cmndcntrlcyber/rtpi --skill hunt-llm-ai-cmndcntrlcyber
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hunt-llm-ai
Source: https://github.com/cmndcntrlcyber/rtpi/tree/main/knowledge_seed/bug_hunter_skills/hunt-llm-ai
Command: npx skills add https://github.com/cmndcntrlcyber/rtpi --skill hunt-llm-ai-cmndcntrlcyber

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill unit is designed to detect and exploit vulnerabilities in AI and LLM systems, focusing on security issues like prompt injection, exfiltration, and agent misuse.

Core Features & Use Cases

  • Prompt Injection: Detects and mitigates direct and indirect prompt injection attacks.
  • Exfiltration: Identifies exfiltration vectors such as ASCII smuggling and tool misuse.
  • Agent Misuse: Monitors and prevents malicious behavior in AI agents.
  • Use Case: When auditing AI systems for security, this Skill can be used to uncover potential vulnerabilities in AI chatbots, RAG endpoints, and agentic systems.

Quick Start

Run the 'hunt-llm-ai' skill to begin the AI/LLM vulnerability assessment.

Frequently Asked Questions about hunt-llm-ai

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

FAQPage Schema
How do I detect prompt injection vulnerabilities in AI chatbots?▼

To detect prompt injection vulnerabilities in AI chatbots, you analyze system prompts and data handling to identify direct and indirect injection vectors. This uncovers malicious behavior risks in agentic systems.

What is ASCII smuggling in LLM security and how does it cause exfiltration?▼

ASCII smuggling in LLM security is an exfiltration vector where attackers misuse tool usage to stealthily extract data. Identifying it requires monitoring AI agents for malicious behavior and analyzing tool usage.

Can I audit RAG endpoints for agent misuse and system prompt leaks?▼

Yes, you can audit RAG endpoints for agent misuse and system prompt leaks. The assessment focuses on analyzing tool usage, data handling, and system prompts to prevent malicious behavior in AI agents.

How do I run a security audit for LLM vulnerabilities in agentic systems?▼

Running a security audit for LLM vulnerabilities in agentic systems involves analyzing system prompts, tool usage, and data handling. This detects prompt injection, exfiltration, and agent misuse across AI chatbots.

What are the limitations of AI security assessments for RAG endpoints?▼

Limitations of AI security assessments for RAG endpoints include the complexity of analyzing dynamic tool usage and data handling. Effective mitigation requires thorough system prompt analysis to catch indirect prompt injection.

Related Skills

cybersecwomancybersecwoman
5

hunt-llm-ai

Detect and mitigate vulnerabilities in LLM/AI systems.

Community
Advanced
tianzecntianzecn
2

working-with-claude-code

Retrieve official Claude Code documentation for configuration, plugins, MCP servers, hooks, and skills.

Community
Advanced
altrupetsaltrupets

Container Security

Automate container security with image scanning, runtime monitoring, and signing.

Official
Advanced
XerrionXerrion

pentest-ai-redteam

Identify and document LLM prompt-injection risks and tool-call abuse in AI systems.

Community
Advanced