Gate on all outgoing text

The idea

Commits, PR text, comments, messages and reports go through rule-based checks (length, structure, banned phrases, attribution, secrets) and a style check against examples a person approved. Failing text gets a limited number of rewrites.

The problem
Workers can leak secrets or send people text in the wrong voice.
Why it matters
No text goes out without the same checks, and the gateway refuses secrets.
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.

  • Learning from the person's own AI history

    A person imports their own transcripts, files, commits and messages. Traits they accept steer briefs, checks and voice. History they bind to one repository becomes facts, decisions and documents that they confirm.

  • One governed tool gateway

    Every action goes through one gateway. It checks authorization, policy, risk, budget and approval, writes the action ledger first, and reads back each external write. A model answer can only add a requirement.

  • Redaction on write and the credential vault

    Secret-shaped values are removed before events, steps, tool results and model context are stored, and customer credentials are sealed with AES-256-GCM bound to org and credential.

  • Only the operator's own words instruct

    Only a verified sender's own lines are instructions; forwarded, quoted, fetched, partner and guest-command text is data, hidden characters are stripped, and actions after instruction-shaped content wait for a person.

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.