← Glossary Company
Zenity
A security company focused on AI agents and low-code business platforms, whose research team publishes flaws in how enterprise AI tools handle permissions.
Zenity works on a corner of security that barely existed a few years ago: the automations, copilots and agents that ordinary employees build inside business software without going through IT. Its research arm, Zenity Labs, publishes the resulting failures.
Its best-known 2026 finding showed how a single tampered ChatGPT link could set up an autonomous agent inside a company, running with the permissions of whoever clicked it. The point of that class of research is that the attack does not break anything technical, it just uses features exactly as designed.
Mentioned in