The store's rules now sit on a Blobs layer: a folder, as before, or a
prefix in an S3-compatible bucket (MinIO locally), chosen by STORE_BACKEND.
The layout is the same in both; the record is written last.
- S3_BUCKET, S3_PREFIX, S3_ENDPOINT_URL, S3_REGION; AWS credentials
- an s3 extra (boto3), installed in the image; moto for tests
- every store test runs on both kinds of storage
Retention, the platform's sign-in and each kept file as a platform
artifact are left: the last needs the platform's S3 backend, which it does
not yet have.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A request names work the Orchestration engine kept instead of carrying its
files: a record or latest link from an earlier document of links (checked,
and of the request's tenant and Customer case), latest/<work>, or an
Execution's id. The kept files and the record that the work completed are
the agent's sources. Nothing starts, and the caller is told why, when the
reference does not resolve to completed work it may use.
A sequence's next agent is given what was kept of the one before, the same
way.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Orchestration engine keeps every completed agent Execution it relays,
by tenant, Customer case and Execution, and serves its record and files at
signed, expiring links. A KAB workflow step, which takes one output per
step, is given one JSON document of links (`deliver: links`); other callers
still get each file, and what they are given is kept either way.
- store.py, links.py, delivery.py, serving.py; the relay keeps and links
- the card declares `deliver`, `sources` and an output schema
- image: STORE_DIR on a /data volume; PUBLIC_URL and LINK_SECRET settings
- option 1's tests replaced by tests of the links document
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>