llm-security

Identify and mitigate LLM security risks using OWASP Top 10 for LLM 2025.

706|30|Updated Mar 13, 2023

Search Tags:#rag-security #prompt-injection #data-leakage #llm-security #system-prompt-leakage #secure-llm

Ready to run

npx skills add https://github.com/wimpysworld/nix-config --skill llm-security-wimpysworld

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: llm-security
Source: https://github.com/wimpysworld/nix-config/tree/main/home-manager/_mixins/development/assistants/skills/llm-security
Command: npx skills add https://github.com/wimpysworld/nix-config --skill llm-security-wimpysworld

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Security guidelines for LLM applications based on OWASP Top 10 for LLM 2025, helping teams design, review, and operate safe AI systems.

Core Features & Use Cases

Proactive risk guidance for building secure LLM apps, including prompt-injection defenses, system prompt leakage mitigation, data leakage controls, access management, rate limiting, auditing, and RAG safeguards. Reactive guidance for questions about security patterns, threat modeling, and secure deployment. Use cases cover chatbots, AI agents, RAG pipelines, and tool-using LLMs across industries.

Quick Start

Apply OWASP Top 10 for LLM 2025 security controls to your current LLM project.

Frequently Asked Questions about llm-security

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

FAQPage Schema

How do I prevent prompt injection attacks in LLM applications?▼

Mitigate system prompt leakage in LLM apps by enforcing output sanitization, access controls, and specific defensive patterns that protect system instructions from being extracted by users.

What security controls do I need for RAG systems?▼

Secure RAG systems require access control, input validation, and data leakage safeguards to prevent unauthorized data retrieval, aligned with OWASP Top 10 for LLM 2025 defensive patterns.

How do I apply OWASP Top 10 for LLM 2025 to my chatbot?▼

Apply OWASP Top 10 for LLM 2025 to chatbots by implementing proactive guardrails including rate limiting, auditing, and access management to identify and mitigate security risks.

Can I use this to stop data leakage in AI agents?▼

Yes, you can stop data leakage in AI agents by applying concrete defensive patterns and proactive guardrails for access management, input validation, and output sanitization.

What is the best way to secure tool-using LLMs against threats?▼

The best way to secure tool-using LLMs is implementing threat modeling, access control, and rate limiting guided by OWASP Top 10 for LLM 2025 to mitigate identified security risks.