make agent-status-token requests --duration=12h and prints the real expiry from the token's exp claim
(tools/agent-status/token_expiry.py, never the token), since the API server may cap a duration.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 943911@bnt-lap001
Assistant-Session: fc322940-1c8e-4bed-ab39-86cec99f23a0
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 943911@bnt-lap001
Assistant-Session: fc322940-1c8e-4bed-ab39-86cec99f23a0
tools/conventions/lint.py (make conventions-lint, contract suite): server-side-apply, one-wait-condition,
one-cluster-route, tool-numbered-receipts, managing-controller (Argo-owned paths from railiance-platform),
with reasoned exceptions. 21 runbook applies made server-side. portal-client-rollout.py (and so
keycape-client.py and the rollout tools) and identity-provisioner-reconcile.py use NK_KUBECONFIG,
default ~/.kube/config-railiance01, without a second kubeconfig or --server override.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 943911@bnt-lap001
Assistant-Session: fc322940-1c8e-4bed-ab39-86cec99f23a0
contracts/owner-pins.json pins six public owners; tools/contract-tests/fetch_owners.py partial-clones
them anonymously (pinned on push, main nightly). tools/contract-tests/run.sh is the one suite list
(make contract-test, CI) and fails when any test skips. New: readiness against real owner source,
replaying both 2026-11-04 states from informed-decision history. pytest pinned by PyPI-checked hashes.
test_approval_request no longer hard-codes /home/worsch/informed-decision.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 943911@bnt-lap001
Assistant-Session: fc322940-1c8e-4bed-ab39-86cec99f23a0
bx ready / request / run / status (close for emergencies): inputs in a state file, receipts numbered
by the tool, every command safe to repeat and one value-free line. OpenBao steps run as silent
children of the warden platform-admin lane (executor-lanes.py --quiet); run closes the window in the
same child, also on error. Runbook sections 5-6 replaced; pilot-run.sh deleted. 134 tests.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 943911@bnt-lap001
Assistant-Session: fc322940-1c8e-4bed-ab39-86cec99f23a0
tools/bounded-execution-readiness/ready.py (make bounded-execution-ready): agent-identity checks from
owner source (admission, approver chain at the Argo pin and main, access-engine profile, CCRs by the
owner's RUNBOOK_ALLOWED_STATUSES, case and broker type pin), and three named read-only founder checks.
executor-lanes.py status: write role, fence, cas_required; no write. A replay test catches 2026-11-04.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 943911@bnt-lap001
Assistant-Session: fc322940-1c8e-4bed-ab39-86cec99f23a0
write-role-audit.py reads the file audit log on stdin and reports each interval in which
the write role existed, with logins, write-policy requests and failed kubernetes logins.
The pilot's own login is the positive control. Receipt holds times, counts and SA names only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 943911@bnt-lap001
Assistant-Session: fc322940-1c8e-4bed-ab39-86cec99f23a0
- canon/schemas/execution-attribution-receipt_v0.1.schema.json: the bounded-executor profile
of the PCC v0.2 receipt, both audit-core events (started, receipt), every object closed.
- tools/execution-attribution-receipt: EAR-1..EAR-10 validator, fixtures generated from the
real executor (completed, reverted, completed-on-retry) with a drift check.
- executor: started carries receipt_version, digests, steps_planned and data.window.not_after
(audit-core AUDIT-WP-0018-T03 request); receipt carries receipt_version and initiator.iss
(configured approver_issuer, else unknown).
- bounded-executor/exercise_audit_sender.py: real executor and client against audit-core's
real ingestion and the Git scope overlay: 202, duplicate 200, conflict 409 as a defect,
wrong source/tenant/token refused and not stored, a receipt-less start flagged by its own
window.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 523928@bnt-lap001
Assistant-Session: 6c737afb-9c14-4152-9f02-477214b9c350
tools/service-binds/executor-lanes.py (bind-lane.py shape): under both
sha256-bound, approved, binding-confirmed CCRs (0058 write, 0059 journey) and
an attended platform-admin, apply both policies, set cas_required on the
svc-authelia key, create the standing journey-read role, refuse if the
window-scoped write role exists, prove the agent fence; verify proves the
journey token reads only its path (<=15m) and denies wrong SA, wrong audience
and write-role login. Value-free receipts. 12 tests on a fake OpenBao (real
helpers), in make bounded-executor-test. Founder runbook:
sso-mfa/k8s/bounded-executor/T10-LANES-RUNBOOK.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 5502@bnt-lap001
Assistant-Session: 68b3b02f-5d0f-4e71-8421-796f62e50196
Consumes one admission once through the approval-engine carrier after an
effect-free preflight (validity, source pins, trust, window vs budget,
binding digest, actor, human aal2 approvers), emits run.started, runs steps
with settle-timed checks, reverts on failure with per-revert checks, retries
within attempts, stops on unknown outcomes, honours budget and kill switch,
and emits one value-free receipt (PDP not-evaluated). 22 tests with fakes.
BEA-9 extended to revert steps; examples regenerated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 994874@bnt-lap001
Assistant-Session: 6017c416-d619-47eb-bd1a-f670a02d7d87
Executor credentials declared (planned) with executor_subject; service-identity
gains convention executor (SI-15: approved-only grants, no person, not
interactive) and --admissions, checking that an admission's ceiling lies inside
the executor's declared grants (pilot covered). RBAC in sso-mfa/k8s/bounded-executor
(not applied): ServiceAccount without automount, a Role limited to the authelia
Deployment and ExternalSecret, no read of credential objects; tests tie Role,
declaration and pilot ceiling. The pilot example's delivery check now reads the
ExternalSecret status. railiance-apps and glas-harness answers recorded (T02).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 994874@bnt-lap001
Assistant-Session: 6017c416-d619-47eb-bd1a-f670a02d7d87
Standard, schema and validator for the founder's admission of one bounded
executor run: exact plan by digest, window by risk class, SI MODE @ SCOPE
ceiling with least-privilege coverage, blast radius by consumer, named
source-pinned checks, revert per mutating step, budget, single use, and the
NK-WP-0040 receipt. check-approval verifies a carrier record (Approval
Engine or founder SSH signature): aal2 human approver, not the executor,
consumed at most once in the window. Examples express the bind split;
step 10 is refused as human-intrinsic. 22 tests; make targets; tool index.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 994874@bnt-lap001
Assistant-Session: 6017c416-d619-47eb-bd1a-f670a02d7d87
prepare-custody, plan, accounts and verify for ld-admin and the svc-* accounts;
named rotations for the Authelia bind and LLDAP_LDAP_USER_PASS.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 308701@bnt-lap001
Assistant-Session: 5548e447-444c-4dcc-bed1-05c16eebfe56
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 308701@bnt-lap001
Assistant-Session: 5548e447-444c-4dcc-bed1-05c16eebfe56
Review against security-canon (Mode of Authority, SEC-DSM 1.0.0) showed the draft
restated canon rules and invented a parallel taxonomy. Draft-3 imports principal_type
and MODE @ SCOPE grants, treats break-glass as an activation, cites SEC-DSM per rule,
states the estate identities in canon terms and adds SI-15 (accountable automation:
prepare, admit, execute, verify, retain). Founder decisions and rule ids are kept.
For agents: tools/security-tools.json and make security-tools list each security tool,
the question it answers and whether an agent may run it; a test keeps the index
complete. AGENTS.md and CLAUDE.md carry the same quickstart block. Validator migration
to grants is T03b. Review record in history/.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 328925@bnt-lap001
Assistant-Session: 14b795fd-c632-4c01-8d28-e7e490c7b0ed
Schema, validator (30 tests), make targets and eight declarations authored from the
inventory, all unconfirmed by their owners. First run: 0 non-conformant, 12 unknown,
2 excepted, 5 planned, 6 conformant. Provisional: built against draft-1 before the
founder's review. Building it changed the draft: custody is a list, SI-X1 now covers
SI-9 and NAMING, and the reserved-name rule applies only to credential-holding
identities (open question 7). Unknown is a finding for every class.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: sonnet
Assistant-Process: 328925@bnt-lap001
Assistant-Session: 14b795fd-c632-4c01-8d28-e7e490c7b0ed
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 3775@bnt-lap001
Assistant-Session: cfacb6a5-ad5b-4f79-ac67-053e3b0d34a3
T2: greenfield live proof against a fresh uninitialized OpenBao 2.5.5 —
caught and fixed 'bao operator unseal -' not reading stdin (now
'bao write sys/unseal key=-'); init and reseal-replay paths proven.
T3: attended-ceremony selectable — runbook, non-secret ceremony-record
template + validator, and a lab/production deployment profile that blocks
sops-held-automation in console selection, gates, and the init script.
T4: console gate + evidence flags for auto-unseal-transit (Helm seal stanza
prepared in railiance-platform).
Also: SCOPE.md refreshed to current repo state; adhoc fix for the broken
check-secrets Make target (unescaped $).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Document three init/unseal custody paths; default sops-held-automation for
fast rebuild cycles. Security bootstrap console lists models, blocks planned
attended-ceremony and auto-unseal-transit with hints, and gates init ceremony
on implemented selection. NET-WP-0020 tracks downstream SSH automation.
- Extended computed validation pattern into main gates:
- Added keycape_openbao_client_deployed() (invokes verify-openbao-client.sh for live check).
- Updated 'KeyCape OpenBao client deployed' gate in build_gates to 'done' if metadata or validator succeeds (T08: UI now proves via validation not just manual flag).
- Added validate-keycape-client subparser, dispatch (prints source+live status), and make target.
- Updated printed available actions list to include it.
- Updated T08 workplan section: status done + detailed 2026-06-03 implementation note (extended from 0019 note; covers one key target as example, pattern for others like LLDAP/privacyIDEA/Authelia using existing verify-*.sh).
- T07 tests + console-test cover; console status gates now reflect more validator output.
- Pragmatic: progress log with task_id, file notes, commit.
- Brief/fix next (expect 8/9 done).
This fulfills T08: more gates compute from validators (ok/fail) rather than manual only; live setup can satisfy checks via the integrated commands.
