Prompt Injection

prompt-injectionmanipulationllmcontrol-flow

Attackers manipulate model behavior by injecting malicious prompts, causing the model to ignore instructions, leak data, or perform unintended actions.

Technical Details
Affected Components:
Language Models (LLMs)Orchestration (Control Flow)kc7.3
Impact Level:High
Attack Vectors
  • Direct Prompt Injection: User-supplied input alters model instructions or context. [High]
  • Indirect Prompt Injection: Malicious content from external sources (e.g., web, email) is injected into the prompt context. [High]
  • Goal Hijacking: Manipulating the model to pursue attacker-defined objectives. [Medium]
Impact Analysis
Risk Score: 9/10
Mitigation Categories
References