skills.rest
jayjpatel9717
jayjpatel9717COMMUNITY
Explore all skills by this author →
GitHub

llm-security

Detect and mitigate LLM security vulnerabilities like prompt injection and RAG poisoning.

Updated Jun 3, 2026
Search Tags:#security#ai#llm#jailbreak#prompt injection#model extraction
One-click install
npx skills add https://github.com/jayjpatel9717/kurukshetra_updated --skill llm-security-jayjpatel9717
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llm-security
Source: https://github.com/jayjpatel9717/kurukshetra_updated/tree/main/squads/ai-security/agents/maya/skills/llm-security
Command: npx skills add https://github.com/jayjpatel9717/kurukshetra_updated --skill llm-security-jayjpatel9717

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill provides comprehensive security testing for LLM/AI applications, detecting vulnerabilities like prompt injection, jailbreaking, data memorization, and RAG poisoning.

Core Features & Use Cases

  • Prompt Injection Detection: Identify and mitigate the risk of system instructions being overridden.
  • System Prompt Leakage Detection: Prevent unauthorized access to system prompts containing confidential information.
  • Training Data Memorization Check: Detect if the model regurgitates sensitive training data.
  • RAG Pipeline Security: Identify and address RAG poisoning vectors and embedding weaknesses.
  • Agent Hijacking Detection: Ensure AI agents cannot be used for unintended actions.
  • Insecure Output Handling: Prevent XSS and injection vulnerabilities in LLM output.
  • Model Extraction Check: Detect model extraction risk via API querying patterns.
  • Sensitive Information Disclosure: Identify any unauthorized exposure of sensitive data.
  • Use Case: For instance, use this Skill to test a chatbot for potential security vulnerabilities and ensure compliance with security best practices.

Quick Start

Run the llm-security skill to test the AI chatbot for prompt injection vulnerabilities.

Frequently Asked Questions about llm-security

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

FAQPage Schema
How do I test my LLM application for prompt injection and jailbreak vulnerabilities?▼

To test for prompt injection and jailbreak vulnerabilities, run a security audit to detect and mitigate system instruction overrides. This identifies risks where user inputs override system prompts.

What is RAG poisoning and how does it affect AI security?▼

RAG poisoning is a security vulnerability where malicious data is injected into retrieval pipelines to manipulate outputs. Security testing identifies RAG poisoning vectors and embedding weaknesses to mitigate this risk.

Can I detect if my LLM is leaking sensitive training data or system prompts?▼

Yes, you can detect training data memorization and system prompt leakage by auditing the model. Security testing identifies if the model regurgitates sensitive training data or exposes confidential system prompts.

How do I prevent model extraction attacks through API querying patterns?▼

Prevent model extraction by analyzing API querying patterns to detect unauthorized data extraction attempts. Security audits identify model extraction risks to protect the intellectual property of your LLM.

How do I secure AI agents against hijacking and insecure output handling?▼

Secure AI agents against hijacking and insecure output handling by testing for unintended actions and injection vulnerabilities. This prevents XSS in LLM output and blocks unauthorized agent actions.

Related Skills

Explore More Software Engineering Skills→
YukiIto1999YukiIto1999

api-security

Automate API security assessment across GraphQL, REST, WebSocket, and Web-LLM interfaces.

Community
Intermediate
cmaennercmaenner
5

agent-security-audit

Audit AI agent configurations for security risks per OWASP LLM Top 10.

Community
Advanced
abhijeetkakade1234abhijeetkakade1234
1

ai-prompt-injection-audit

Identify and remediate prompt injection and jailbreak vulnerabilities in LLM applications.

Community
Advanced
kangarookingkangarooking
154

injection-defense

Detect and block prompt injection, jailbreak, and social engineering attempts.

Community
Advanced