Security review of Parmana

This page is for the person who has to decide whether Parmana is safe to put in front of an AI agent that can act on your business systems. It lists what Parmana checks, what it records, how a third party can verify a record, and what it does not prove.

What Parmana checks before an action runs

Who can change a policy

What is recorded, and who can verify it

Caller type

An AI agent, a person, and a service account go through the same execution path and produce the same record shapes. Only policy authorship is restricted to verified humans.

What the evidence does not prove

Technical detail is in the documentation. Questions about a specific control: see the questions page.