skills.rest
bpnrockstar
bpnrockstarCOMMUNITY
Explore all skills by this author →
GitHub

hunt-llm-ai

Identify and exploit AI system vulnerabilities using prompt injection techniques.

Updated Jul 1, 2026
Search Tags:#vulnerability testing#ai security#prompt injection#red teaming#ai exploitation
One-click install
npx skills add https://github.com/bpnrockstar/UnifiedBugHunter --skill hunt-llm-ai-bpnrockstar
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hunt-llm-ai
Source: https://github.com/bpnrockstar/UnifiedBugHunter/tree/main/skills/hunt-llm-ai
Command: npx skills add https://github.com/bpnrockstar/UnifiedBugHunter --skill hunt-llm-ai-bpnrockstar

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires llm_redteam.py, llm_payloads, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill helps identify and exploit vulnerabilities in AI systems through sophisticated prompt injection techniques, ensuring security and reliability.

Core Features & Use Cases

  • Advanced Prompt Injection: Utilizes 15+ advanced techniques to detect and exploit AI vulnerabilities.
  • Multi-Turn Escalation: Exploits the erosion of safety across conversation turns for deeper security testing.
  • Exfiltration Channels: Identifies and validates exfiltration channels, including OOB callbacks and system prompt leakage.
  • Automated Red Teaming: Provides an automated tool for testing LLM endpoints and payloads.
  • MCP Server Security Testing: Tests for tool discovery abuse, file system access, database injection, network SSRF, code execution, supply chain, and tool poisoning.
  • Use Case: For security professionals looking to identify and fix AI vulnerabilities in chatbots, RAG systems, and other AI-driven applications.

Quick Start

Run the hunt-llm-ai skill with the target URL to start testing for AI vulnerabilities.

Frequently Asked Questions about hunt-llm-ai

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

FAQPage Schema
How do I test my chatbot for prompt injection vulnerabilities?▼

You can test chatbots for prompt injection vulnerabilities by running automated red teaming tools against the target URL to identify and exploit AI system weaknesses using advanced multi-turn escalation techniques.

What is multi-turn escalation in AI red teaming?▼

Multi-turn escalation in AI red teaming is a testing technique that exploits the erosion of safety guardrails across conversation turns to achieve deeper system access and validate potential security weaknesses.

How do I check for system prompt leakage and exfiltration channels?▼

You can check for system prompt leakage by validating exfiltration channels, including out-of-band callbacks, through automated payload testing designed to expose hidden system instructions and data pathways.

Can I use automated red teaming to test MCP server security?▼

Yes, automated red teaming supports MCP server security testing by identifying tool discovery abuse, file system access, database injection, network SSRF, code execution, supply chain, and tool poisoning vulnerabilities.

Does this AI security testing tool work with RAG systems?▼

Yes, this AI security testing tool works with RAG systems, targeting retrieval-augmented generation applications alongside chatbots and other AI-driven platforms to identify and fix exploitable vulnerabilities.

What are the limitations of automated LLM vulnerability testing?▼

Automated LLM vulnerability testing primarily focuses on prompt injection, exfiltration validation, and MCP server abuse, meaning it may not cover non-prompt-based architectural flaws or logic bugs unrelated to text manipulation.

Related Skills

naporin0624naporin0624
3

playwright-security-runner

Execute XSS and SQL injection payloads against web forms with Playwright.

Community
Advanced
uphiagouphiago
1.1k

hunt-llm-ai

Identify exploitable LLM/AI vulnerabilities crossing trust boundaries with validation gates.

Community
Advanced
YukiIto1999YukiIto1999

performing-blind-ssrf-exploitation

Identify and exploit blind SSRF vulnerabilities using out-of-band callbacks and timing analysis.

Community
Advanced
aivos-xieaivos-xie
1

web-security

Scan web applications for vulnerabilities and exploit them with sqlmap, Nikto, nuclei, and Metasploit.

Community
Advanced