Round-10 review (railiance-platform) found that ContainedOperator inherited
an ssh-only kubectl, so an E2 would write OpenBao and then fail its sync,
restart and rollback. Also: a bad AUDIT_KUBECONFIG now leaves window-error
evidence, and the receipt script validates bodies and lets a later
request-changes withdraw an approval.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 236506@bnt-lap001
Assistant-Session: 09087018-91e1-48c5-ab2f-acfbb62c92e4
Since RCLUSTER-WP-0019 the node account cannot read the k3s admin kubeconfig,
so ssh railiance01 kubectl fails at read node state. With AUDIT_KUBECONFIG
set, cluster reads and the port-forward run locally against that file; host
reads stay on ssh. Unset keeps the earlier path. The E2 receipt script takes
owner approvals for the current contract digest from the hub.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 236506@bnt-lap001
Assistant-Session: 09087018-91e1-48c5-ab2f-acfbb62c92e4
One contained login: project one synthetic sender, event accepted, replay
must be duplicate (exactly once), cleanup; no lease is touched.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 680842@bnt-lap001
Assistant-Session: 54eb9e8e-7d6b-4f0c-b7a6-a6db62075883
Harness used the retracted public OpenBao URL and a token file; lane used a
standing token file. Both now run only inside the contained ops-warden login
(tunnel, helper token), and t02-attended-window.py runs project, exercise and
cleanup under one login without printing (warden fails closed on child output).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 680842@bnt-lap001
Assistant-Session: 54eb9e8e-7d6b-4f0c-b7a6-a6db62075883
