Lethal trifecta
Combination, in one AI agent, of access to private data, exposure to untrusted content and the ability to communicate externally, allowing an attacker to trick the agent into sending private data to them.
In plain language
An agent that reads untrusted content, reaches private data and can send messages out can be tricked into leaking that data.
In a sentence
- The security review checked whether the agent had the lethal trifecta of capabilities.
- The team discussed the lethal trifecta before connecting private records to an agent that could send messages.
Usage note
The phrase names a combination of capabilities that exposes private data to theft. It does not establish that a particular agent has been compromised.
See also
- YOLO mode
Mode of running an AI agent with every action approved automatically, so its tool calls execute without asking the user for permission first.
- AI control
Approach to AI safety that aims to keep a model unable to cause damage even if it is badly misaligned and tries to subvert its safeguards.
Cite this term.
CC BY 4.0, attribution required.
Orbyt Labs, Lethal trifecta (AI Glossary), Nov 14, 2026. https://
Orbyt Labs Glossaries (Nov 14, 2026), https://
Last updated .