Indirect Prompt Injection (T1698)

Tactic: Execution · Platforms: Windows, SaaS

The interactive view maps 6 mitigations to this technique, alongside D3FEND countermeasures and data-component coverage.

Overview

Adversaries may deliver malicious prompts to an LLM indirectly by embedding the prompts in content that the adversary expects may be processed by an LLM. These prompts can be included in web pages, documents, emails, tickets, code, or other sources that an agent, copilot, or other AI system may retrieve and process. When an LLM processes this content, the embedded instructions may cause the system to perform unintended behaviors.

Adversaries may use indirect prompt injection to Impair Autonomous Defenses, and for unauthorized execution of any capabilities available to the LLM.

Explore

• Open T1698 in the interactive ATT&CK matrix
• ATT&CK Intelligence View guide
• Official MITRE ATT&CK entry for T1698
• Sigma detection rules for MITRE ATT&CK
• Built by Shahrukh Khan