Portable learned state

The idea

An owner exports preferences, rule candidates, checkers, approved examples and procedures as inert archives. An import creates private copies and leaves out other people's data.

The problem
Learned state was locked in one organization.
Why it matters
What the crew learned can move without leaking private data.
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.

  • State in the repository (.agents/)

    A repository's .agents/ folder holds rules, memory notes, task checks and summaries. It is shared through git and read back in.

  • Forget reaches derived copies

    When a person forgets a note, the platform also rewrites the copies that work made from it: briefs, plans, findings, logs and outputs. It also cancels approvals and holds outreach that quote the note.

  • Stated data retention

    Settings shows what is kept and for how long from the constants in force; full model inputs are kept 7 days and only the requester can read or delete them.

  • Hybrid memory recall

    Recall ranks memories with full-text and trigram search, plus vectors where embeddings exist. It says which arms ran and why vectors are off, and it gives the worker its readable memory as files.

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.