Prompt manipulation techniques to induce language models and generative AI systems to produce unintended outputs, bypass security filters, or reveal sensitive information. Includes prompt injection, jailbreaking, context manipulation, and evasion strategies against guardrails implemented in Large Language Models.