ai-safe2-framework

Agency Check integration

Agency Check is a human decision interface built on the same evidence discipline as AI SAFE2. It is not a software merge gate.

The service planned for agency-check.app should translate a bounded evidence set into one Decision Card:

Every card should preserve:

It must not call an action safe, execute the action, or convert an absence of evidence into approval.

Public reference: https://cyberstrategyinstitute.com/agency-check/