Glossary · Term

Oracle Poisoning

← all terms

Definition

An attack where the database an AI agent consults gets quietly corrupted, so the agent confidently reasons from false facts.

An attack class targeting structured data sources (knowledge graphs, MCP-served databases) that LLM agents trust as observational ground truth, corrupting grounding without altering prompts, training, or tool behavior.

Mentioned in 1 episode

  1. 039
    When Smarter Agents Get Fooled by Three Extra Nodes in a Database