What it does
A Python library that checks each action an AI agent wants to run against rules written in YAML. Allowed actions go through, sensitive ones wait in an approval queue for a person, and forbidden ones are blocked. Every action gets a receipt linked to the one before it, and the data can live in a single file or in Postgres.
Use cases
- 01Require approval before a sensitive action runs
- 02Block actions that the policy forbids
- 03Keep a receipt trail of every approved action