Structured and plain-language policies

The idea

Policies with worker > role > team > org precedence gate actions; a one-sentence rule drafts a policy with a preview, and free-text policies are read at the gateway where a model can only add an approval.

The problem
Business rules lived in people's heads or needed structured editing.
Why it matters
An owner writes a rule in plain words and sees which past actions it would match.
Where it stands
In production as of

A step-by-step replay of this idea is not published yet. The library adds one when a recorded run shows it.

Where it lives in the product

  • Autonomy modes and ask-for list

    An org or worker picks Ask first, Auto or Full autonomy; risk tiers and a fixed ask-for list (merge, deploy, outside speech, spend, hiring, reorg, new integration) decide what waits for a person.

  • Approvals on web, WhatsApp, Slack and Teams

    A governed action waits for an approval that eligible people decide on the web, WhatsApp buttons or Slack/Teams direct messages; the requester never decides, and owned resources are approved by their owner.

Turn the numbers into shipped work.

Crewshal runs these choices for you: a persistent AI worker with memory and rules, on your Claude and Codex subscriptions.