Record current Yggdrasil body split

notes/fresh-workspace-handoff.md

@@ -9,18 +9,20 @@ OpenAI-compatible authentication and model-transport boundary.
 
 The local `main` branch is clean and synchronized with its upstream.
 
-On Yggdrasil:
+Current host reality:
 
-- Idunn is active.
-- Bifrost is active.
-- Epiphany services are inactive after a clean rollback.
-- The canonical Idunn deployment brake is engaged.
-- A deployment request is awaiting authorization.
-- The current Idunn incarnation did not consume a matching valid grant for that
-  request.
+- The intended bare-metal body at `94.130.223.131` is unreachable on ports 22,
+  80, and 443 from both the operator workstation and the live production VM.
+- DNS, SSH, public applications, and WireGuard currently terminate on
+  `94.130.75.184`.
+- That reachable fallback is a 2 GB VM with 15 GB free. It runs Bifrost and the
+  public application stack, but has no Idunn or Epiphany service body.
+- The operator workstation WireGuard service has been restored, but its
+  `10.77.0.1` peer is also the fallback VM.
+- The Hetzner console is not authenticated in the available browser session.
 
-No deployment, cognition release, broad Wake, or reboot is authorized while
-that mismatch remains unexplained.
+No deployment, cognition release, broad Wake, or reboot is authorized until the
+deployment Body is recovered or deliberately reassigned.
 
 ## Authority map
 
@@ -39,10 +41,11 @@ that mismatch remains unexplained.
 
 ## Next action
 
-Inspect the stable Idunn request, owner, grant, and incarnation ledger without
-mutation. Identify the exact transition that prevented grant consumption.
-Only after that transition is understood should a new deployment decision be
-made.
+Sign in to the Hetzner console and inspect the power, rescue, and network state
+of `94.130.223.131`. Recover it if it remains the intended canonical Body. If it
+cannot be recovered, explicitly decide whether to adopt the constrained fallback
+VM before installing Idunn or Epiphany there. Once the Body is settled, use the
+canonical Idunn deployment path and verify sustained braked readiness.
 
 ## Immediate Re-entry Instruction
 

state/map.yaml

@@ -4,7 +4,7 @@ experiment:
   purpose: Native, inspectable agent organism with typed state and explicit organ authority.
 
 objective:
-  current: Restore a trustworthy braked Yggdrasil deployment, then prove one bounded Persona conversation through the governed Bifrost crossing.
+  current: Restore an authoritative Yggdrasil body, deploy Epiphany through Idunn, and prove sustained braked readiness before one bounded Persona conversation.
 
 architecture:
   body: Repository, runtime, service processes, stores, interfaces, and deployment substrate.
@@ -34,25 +34,28 @@ invariants:
   - Evidence and git retain history; this map retains only state that changes future decisions.
 
 current_status:
-  summary: Epiphany is a native Rust/CultCache/CultMesh/CultNet organism. The local branch is clean and synchronized. Idunn and Bifrost are active on Yggdrasil; Epiphany services are inactive after a clean rollback; the canonical deployment brake is engaged. One deployment request is awaiting authorization, but its grant was not consumed by the current Idunn incarnation. No further deployment or cognition release is authorized until that ownership chain is understood.
+  summary: Epiphany is a native Rust/CultCache/CultMesh/CultNet organism, but no reachable host currently runs its canonical service body. The intended bare-metal Yggdrasil at 94.130.223.131 is unreachable on SSH, HTTP, and HTTPS from both the operator workstation and the live production VM. DNS, public applications, SSH, and WireGuard currently terminate on 94.130.75.184, a 2 GB fallback VM that runs Bifrost and the public stack but has no Idunn or Epiphany services. Provider-console inspection requires operator login. Choosing or recovering the deployment Body now precedes every Idunn request or grant.
   operational_state:
     local_worktree: clean
-    deployment_brake: engaged
-    idunn: active
-    bifrost: active
-    epiphany_services: inactive_after_clean_rollback
-    pending_deployment_request: awaiting_authorization
-  unresolved_authority: Determine why the stable pending deployment request was not matched and consumed by the current Idunn request/owner ledger.
-  next_action: Inspect the Idunn request, owner, grant, and incarnation ledger as read-only state; identify the exact broken authority transition before issuing any grant or release.
+    operator_wireguard: restored
+    intended_bare_metal_body: unreachable
+    live_fallback_body: production_apps_active_without_idunn_or_epiphany
+    deployment_brake: unavailable_with_unreachable_body
+    idunn: not_present_on_reachable_body
+    bifrost: active_on_fallback_body
+    epiphany_services: not_present_on_reachable_body
+  unresolved_authority: Decide whether 94.130.223.131 remains the canonical Body and must be recovered through Hetzner, or whether 94.130.75.184 is deliberately adopted and rebuilt as the canonical deployment Body.
+  next_action: Sign in to the Hetzner console and inspect the power, rescue, and network state of 94.130.223.131; if it cannot be recovered, explicitly authorize adoption of the constrained fallback VM before installing Idunn or Epiphany there.
 
 active_subgoals:
-  - Diagnose the pending Idunn request/owner/grant chain without mutation.
-  - Refresh this map and the handoff immediately when live evidence changes the operational state.
+  - Recover provider sight or obtain an explicit deployment-Body decision.
+  - Deploy Epiphany only after the Body owner is settled, then prove sustained braked readiness.
 
 forbidden_actions:
   - Do not issue a speculative deployment grant.
   - Do not broadly Wake the swarm.
   - Do not reboot Yggdrasil as part of diagnosis.
+  - Do not install the control plane on the fallback VM without explicitly adopting that constrained Body.
   - Do not treat stored status, process liveness, or an old receipt as current authority.
 
 memory_policy: